You are on page 1of 539

United States Department of Agriculture Forest Service Pacific Southwest Forest and Range Experiment Station General Technical

Report PSW110

Proceedings of the CALIFORNIA RIPARIAN SYSTEMS CONFERENCE


September 22-24, 1988 Davis, California
Protection, Management, and Restoration for the 1990's

Abell, Dana L., Technical Coordinator. 1989. Proceedings of the California Riparian Systems Conference: protection, management, and restoration for the 1990s; 1988 September 22-24; Davis, CA. Gen. Tech. Rep. PSW-110. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 544 p. The nearly 100 papers in these proceedings are aimed at a diverse audience of resource managers, environmental consultants, researchers, landowners, environmental activists, and a variety of user groups. Some of the papers explain how streams interact with the plants and animals at their margins and with the land that they occupy to accomplish a range of important functions, including protecting the banks from erosion, reducing the impacts of flooding, providing wildlife habitat, protecting instream habitat for fishes, producing forage for livestock, and enhancing human lives. Biological diversity in western lands is often directly related to these corridors, which also serve as major routes for migratory birds. Special attention is given to the several threatened and endangered species which need riparian habitats and to the response of riparian systems to such disturbances as fire, logging, landslides and diversion for power or water supply. A concluding section deals with measures being taken to preserve and restore riparian lands, particularly along large rivers and in the cities. Special attention is given in some of these papers to revegetation techniques. Retrieval Terms: riparian habitat, riparian systems, biological diversity, revegetation, stream diversion, threatened and endangered species, range management.

Foreword

Authors of papers in these Proceedings provided manuscripts that were edited and returned to them with technical reviewers' comments. Authors were responsible for following up and acting on the reviewers comments. Final manuscripts were processed electronically for publication by using the TeX computer program. The views expressed in each paper are those of the authors and not necessarily those of the sponsoring organizations. Trade names and commercial products or enterprises are mentioned solely for information. No endorsement by any of the sponsoring organizations is implied. Papers may mention pesticides, but this does not mean that the pesticide uses reported are recommended nor that such uses have been registered by the appropriate governmental agencies.

Sponsors

Co-Sponsors

University Extension, University of California, Davis California Department of Water Resources California Reclamation Board California Department of Transportation Bureau of Land Management, U.S. Department of Interior Forest Service, U.S. Department of Agriculture Pacific Gas & Electric Company Southern California Edison Company California Department of Fish and Game California Energy Commission Bureau of Reclamation, U.S. Department of Interior

Corps of Engineers, U.S. Army Soil Conservation Service, U.S. Department of Agriculture Chevron USA Water Resources Center, University of California Institute of Ecology, University of California, Davis National Park Service, U.S. Department of Interior LSA Associates Harvey & Stanley Associates Jones & Stokes Associates Sacramento River Preservation Trust

Publisher:

Pacific Southwest Forest and Range Experiment Station P.O. Box 245, Berkeley, California 94701 June 1989

Proceedings of the CALIFORNIA RIPARIAN SYSTEMS CONFERENCE


September 22-24, 1988 Davis, California

Protection, Management, and Restoration for the 1990's

Contents
PREFACE .............................................................................................................................................................................i

SESSION A: CHANNEL DYNAMICS AND RIPARIAN SYSTEMS Introduction ................................................................................................................................................................1 Chuck Watson and Dana L. Abell Influence of Valley Floor Landforms on Stream Ecosystems ....................................................................................3 Stanley V. Gregory, Gary A. Lamberti, and Kelly M. S. Moore Channel-Dynamic Control on the Establishment of Riparian Trees After Large Floods in Northwestern California ...................................................................................................................9 Thomas E. Lisle Alder Establishment and Channel Dynamics in a Tributary of the South Fork Eel River, Mendocino County, California ........................................................................................................14 William J. Trush, Edward C. Connor, and Alan W. Knight The Middle Sacramento River: Human Impacts on Physical and Ecological Processes Along a Meandering River .......................................................................................................................22 Koll Buer, Dave Forwalter, Mike Kissel, and Bill Stohlert Influence of Channel Geomorphology on Retention of Dissolved and Particulate Matter in a Cascade Mountain Stream ....................................................................................................33 Gary A. Lamberti, Stan V. Gregory, Linda R. Ashkenas, Randall C. Wildman, and Alan G. Steinman A New Approach to Flood Protection Design and Riparian Management ...............................................................40 Philip B. Williams and Mitchell L. Swanson Effects of Bank Revetment on Sacramento River, California ..................................................................................47 Michael D. Harvey and Chester C. Watson Post-Fire Interactions Between Riparian Vegetation and Channel Morphology and the Implications for Stream Channel Rehabilitation Choices ............................................................................51 Susan C. Barro, Peter M. Wohlgemuth, and Allan G. Campbell

Meanderbelt Dynamics of the Sacramento River, California ...................................................................................54 Michael D. Harvey

SESSION B: CENTRAL VALLEY RIPARIAN FORESTS Introduction ..............................................................................................................................................................61 F. Jordan Lang and Dana L. Abell Feasibility of Mapping Riparian Habitats Under Natural Conditions in California .................................................63 David R. Dawdy Great Valley Riparian Habitats and the National Registry of Natural Landmarks ...................................................69 Robert F. Holland and Cynthia L. Roye Plant Community Development, Site Quality Analysis and River Dynamics in the Design of Riparian Preserves on the Middle Sacramento River, California .............................................................74 Niall F. McCarten San Joaquin River Riparian Habitat Below Friant Dam: Preservation and Restoration ...........................................79 Donn Furman Middle Sacramento River Refuge: A Feasibility Study ............................................................................................83 Charles J. Houghten and Frank J. Michny Developing Management Plans for California Riparian Systems .............................................................................88 Michael Josselyn, Molly Martindale, Dianne Kopec, and Joan Duffield

SESSION C: RANGELAND AND DESERT RIPARIAN SYSTEMS Introduction ..............................................................................................................................................................91 John W. Willoughby Rangeland Riparian Systems ....................................................................................................................................93 Wayne Elmore Using Stream Classification to Prioritize Riparian Rehabilitation After Extreme Events ........................................96 Sherman Swanson Ten Years of Change in Sierran Stringer Meadows: An Evaluation of Range Condition Models ................................102 Barbara H. Allen An Application of BLM's Riparian Inventory Procedure to Rangeland Riparian Resources in the Kern and Kaweah River Watersheds ...............................................................................................109 Patricia Gradek, Lawrence Saslaw, and Steven Nelson The Fallacy of Structures and the Fortitude of Vegetation ..........................................................................................................116 Wayne Elmore and Robert L. Beschta Evidence for an Alternative Landscape Potential in California Annual Rangelands .........................................................120 Richard J. King Use of Supplemental Feeding Locations to Manage Cattle Use on Riparian Areas of Hardwood Rangelands .....124 Neil K. McDougald, William E. Frost, and Dennis E. Jones Clark Canyon (Mono County) Riparian Demonstration Area ....................................................................................................127 John W. Key and Mark A. Gish

Southwestern Woody Riparian Vegetation and Succession: An Evolutionary Approach ....................................... 135 R. Roy Johnson, Peter S. Bennett, and Lois Haight Relative Nature of Wetlands: Riparian and Vegetational Considerations ............................................................... 140 Peter S. Bennett, Michael R. Kunzmann, and R. Roy Johnson The Riparianness of a Desert Herpetofauna ............................................................................................................ 143 Charles H. Lowe Coyote Creek (San Diego County) Management and Restoration at Anza-Borrego Desert State Park .................. 149 David H. Van Cleve, Lyann A. Comrack, and Harold A. Wier

SESSION D: RIPARIAN SYSTEMS AND FOREST MANAGEMENT Introduction ............................................................................................................................................................. 155 Bruce J. McGurk Predicting Stream Temperature After Riparian Vegetation Removal ..................................................................... 157 Bruce J. McGurk Coarse Woody Debris Ecology in a Second-Growth Sequoia sempervirens Forest Stream ................................... 165 Matthew D. O'Connor and Robert R. Ziemer Pacific Yew: A Facultative Riparian Conifer with an Uncertain Future ................................................................. 172 Stanley Scher and Bert Schwarzschild Riparian Systems and Forest ManagementChanges in Harvesting Techniques and their Effects on Decomposed Granitic Soils ..................................................................................................... 176 John W. Bramhall Stabilization of Landslides for the Improvement of Aquatic Habitat ...................................................................... 180 Michael J. Furniss

SESSION E: COASTAL STREAMSECOLOGY AND RECOVERY Introduction ............................................................................................................................................................. 185 Earle W. Cummings Alluvial Scrub Vegetation in Coastal Southern California ...................................................................................... 187 Ted L. Hanes, Richard D. Friesen, and Kathy Keane Recovery of the Chaparral Riparian Zone After Wildfire ....................................................................................... 194 Frank W. Davis, Edward A. Keller, Anuja Parikh, and Joan Florsheim Riparian Restoration and Watershed Management: Some Examples from the California Coast ............................ 204 Laurel Marcus Restoring and Maintaining Riparian Habitat on Private Pastureland ...................................................................... 211 Nancy Reichard Recovery of Riparian Vegetation on an Intermittent Stream Following Removal of Cattle ................................... 217 Jerry J. Smith Giant Reed (Arundo donax): A Climax Community of the Riparian Zone ............................................................. 222 John P. Rieger and D. Ann Kreager

Techniques for Minimizing and Monitoring the Impact of Pipeline Construction on Coastal Streams .................. 226 Thomas W. Mulroy, John R. Storrer, Vincent J. Semonsen, and Michael L. Dungan

SESSION F: WILDLIFE IMANAGING FOR SELECTED SPECIES Introduction ............................................................................................................................................................. 233 Peter B. Moyle Stonefly (Plecoptera) Feeding Modes: Variation Along a California River Continuum ........................................ 235 Richard L. Bottorff and Allen W. Knight Habitat and Populations of the Valley Elderberry Longhorn Beetle Along the Sacramento River ......................... 242 F. Jordan Lang, James D. Jokerst, and Gregory E. Sutter Practical Techniques for Valley Elderberry Longhorn Beetle Mitigation ............................................................... 248 Greg Sutter, Jeurel Singleton, Jim King, and Ann Fisher How Tight is the Linkage Between Trees and Trout? ............................................................................................. 250 Margaret A. Wilzbach Geomorphic and Riparian Influences on the Distribution and Abundance of Salmonids in a Cascade Mountain Stream .......................................................................................................... 256 Kelly M.S. Moore and Stan V. Gregory Montane Riparian Habitat and Willow Flycatchers: Threats to a Sensitive Environment and Species ................... 262 Susan D. Sanders and Mary Anne Flett Population Trends and Management of the Bank Swallow (Riparia riparia) on the Sacramento River, California .................................................................................................................................. 267 Barrett A. Garrison, Ronald W. Schlorff Joan M. Humphrey, Stephen A. Laymon, and Frank J. Michny A Proposed Habitat Management Plan for Yellow-Billed Cuckoos in California ................................................... 272 Stephen A. Laymon and Mary D. Halterman Characteristics of the Least Bell's Vireo Nest Sites Along the Santa Ynez River, Santa Barbara County .............................................................................................................................................. 278 Thomas E. Olson and M. Violet Gray Description of Nesting Habitat for Least Bell's Vireo in San Diego County ........................................................... 285 Bonnie J. Hendricks and John P. Rieger Maintaining Site Integrity for Breeding Least Bell's Vireos .................................................................................... 293 James M. Greaves Use of Non-Riparian Habitats by Least Bell's Vireos ............................................................................................. 299 Barbara E. Kus and Karen L. Miner

SESSION G: WILDLIFE IIMANAGING WILDLIFE ASSOCIATIONS WITHIN RIPARIAN SYSTEMS Introduction ............................................................................................................................................................. 305 Jo Anne Sorenson A Test of the California Wildlife-Habitat Relationship System for Breeding Birds in Valley-Foothill Riparian Habitat ........................................................................................................................ 307 Stephen A. Laymon

Avifauna and Riparian Vegetation in Carmel Valley, Monterey County, California .............................................. 314 Molly Williams and John G. Williams Wildlife Monitoring of a Riparian Mitigation Site : ................................................................................................. 319 Michael Rigney, L. Richard Mewaldt, Blair O. Wolf, and Ronald R. Duke Status Changes of Bird Species Using Revegetated Riparian Habitats on the Lower Colorado River from 1977 to 1984 .......................................................................................................................... 325 Bertin W. Anderson, William C. Hunter, and Robert D. Ohmart Bird Use of Natural and Recently Revegetated Cottonwood-Willow Habitats on the Kern River .......................... 332 William C. Hunter, Bertin W. Anderson, and Reed E. Tollefson The Upper Santa Ynez River as Habitat for a Diverse Riparian Flora and Fauna ................................................... 339 M. Violet Gray, James M. Greaves, and Thomas E. Olson Activities and Ecological Role of Adult Aquatic Insects in the Riparian Zone of Streams ..................................... 342 John K. Jackson and Vincent H. Resh

SESSION H: EFFECTS OF STREAM DIVERSIONS ON CALIFORNIA RIPARIAN SYSTEMS Introduction ............................................................................................................................................................. 347 Roland J. Risser and Carl A. Fox Hydrology of Bishop Creek, California: An Isotopic Analysis ............................................................................... 349 Michael L. Space, John W. Hess, and Stanley D. Smith Stream-Groundwater Interactions Along Streams of the Eastern Sierra Nevada, California: Implications for Assessing Potential Impacts of Flow Diversions .......................................................................... 352 G. Mathias Kondolf Water Relations of Obligate Riparian Plants as a Function of Streamflow Diversion on the Bishop Creek Watershed ......................................................................................................................................... 360 Stanley D. Smith, Janet L. Nachlinger, A. Bruce Wellington, and Carl A. Fox Riparian Plant Water Relations Along the North Fork Kings River, California ...................................................... 366 Janet L. Nachlinger, Stanley D. Smith, and Roland J. Risser A Riparian Vegetation Ecophysiological Response Model ..................................................................................... 370 Jeffrey P. Leighton and Roland J. Risser Water Relations of White Alder .............................................................................................................................. 375 Virginia I. Dains Interpreting Physiological Data from Riparian Vegetation: Cautions and Complications ...................................... 381 John G. Williams Riparian Vegetation Base-line Analysis and Monitoring Along Bishop Creek, California ..................................... 387 Janet L. Nachlinger, Carl A. Fox, and Patricia A. Moen Riparian Communities of the Sierra Nevada and their Environmental Relationships ............................................. 393 Richard R. Harris Early Recovery of an Eastern Sierra Nevada Riparian System After 40 Years of Stream Diversion ...................... 399 Julie C. Stromberg and Duncan T. Patten

The Effect of Water Management and Land Use Practices on the Restoration of Lee Vining and Rush Creeks ...................................................................................................................................................... 405 Peter Vorster and G. Mathias Kondolf

SESSION I: IMPLEMENTING REVEGETATION PROJECTS Introduction ............................................................................................................................................................. 411 John T. Stanley Research as an Integral Part of Revegetation Projects ............................................................................................. 413 Bertin W. Anderson Juniper for Streambank Stabilization in Eastern Oregon ......................................................................................... 420 Guy R. Sheeter and Errol W. Claire A Low Cost Brush Deflection System for Bank Stabilization and Revegetation .................................................... 424 Mary Elizabeth Meyer Reestablishment of Native Riparian Species at an Altered High Elevation Site ..................................................... 428 Franklin J. Chan and Raymond W. Wong Watershed Restoration in the Northern Sierra Nevada: A Biotechnical Approach ................................................. 436 Donna S. Lindquist and Linton Y. Bowie Revegetation of Riparian Trees and Shrubs on Alluvial Soils Along the Upper Sacramento River, 1987-1988 ..... 441 Steven P. Chainey, F. Jordan Lang, and Skip Mills Coyote Creek (Santa Clara County) Pilot Revegetation Project .............................................................................. 447 John T. Stanley, L. R. Silva, H. C. Appleton, M. S. Marangio, W. J. Lapaz, and B. H. Goldner Revegetation Along Coyote Creek (Santa Clara County) at Two Freeway Bridges ............................................... 455 Veda L. Lewis and Keith A. Robinson The Crescent Bypass: A Riparian Restoration Project on the Kings River (Fresno County) .................................. 457 Jonathan A. Oldham and Bradley E. Valentine A Restoration Design for Least Bell's Vireo Habitat in San Diego County ............................................................ 462 Kathryn J. Baird and John P. Rieger Creating Habitat for the Yellow-Billed Cuckoo (Coccyzus americana) ................................................................. 468 Bertin W. Anderson and Stephen A. Laymon Initial Development of Riparian and Marsh Vegetation on Dredged-material Islands in the Sacramento-San Joaquin River Delta, California .......................................................................................... 473 A. Sidney England, Mark K. Sogge, and Roy A. Woodward Air-Earth Interface Model for Restoring Riparian Habitats .................................................................................... 476 Robert M. Dixon

SESSION J: URBAN STREAMS Introduction ............................................................................................................................................................. 483 A. L. Riley

The Wildcat-San Pablo Creek Flood Control Project and Its Implications for the Design of Environmentally Sensitive Flood Management Plans ......................................................................................... 485 A. L. Riley Riparian and Related Values Associated with Flood Control Project Alternatives at Wildcat and San Pablo Creeks ............................................................................................................................. 491 Philip A. Meyer Redesign of a Flood Control Project by Citizen Initiative ....................................................................................... 495 Bev Ortiz Innovations in Stream Restoration and Flood Control Design Meeting Flood Capacity and Environmental Goals on San Luis Obispo Creek .............................................................................. 501 Wayne Peterson Public Participation and Natural Habitat Preservation Along Arcade Creek, Del Paso Regional Park, Sacramento, California ............................................................................................................................................ 506 Timothy J. Vendlinski and Steven N. Talley Arroyo Management Plan (Alameda County): A Plan for Implementing Access and Restoring Riparian Habitats ..................................................................................................................................... 512 Kent E. Watson, Jim Horner, and Louise Mozingo

SESSION K: COORDINATING INTEREST GROUPS Introduction ............................................................................................................................................................. 519 Dana L. Abell Conflicts in River Management: A Conservationist's Perspective on Sacramento River Riparian HabitatsImpacts, Threats, Remedies, Opportunities, and Consensus .......................................... 521 Richard Spotts Riparian Area Management: Principles, Politics, and Practices .............................................................................. 526 John W. Ross and Sheila L. Massey Integrated Riparian Area Management on the Tule Lake Allotment, Lassen County ............................................. 530 Bill Flournoy, Don Lancaster, and Paul Roush Riparian Protection Rules for Oregon Forests ......................................................................................................... 533 George G. Ice, Robert L. Beschta, Raymond S. Craig, and James R. Sedell Formation of the Arizona Riparian Council: An Example of Lasting Public Interest in Riparian Resources ......... 537 Duncan T. Patten and William C. Hunter

APPENDIX: Riparian Conference Advisory Committee ........................................................................................... 541 AUTHOR INDEX .......................................................................................................................................................... 543

PREFACE
This volume presents the proceedings of the second large conference to be convened at the University of California, Davis, under the California Riparian Systems title. It is one of the many responses since the first expression of public concern in the mid-1970's over the catastrophic loss of these attractive and valuable streamside lands. By the time of the first big California riparian conference in 1981, the concern had already been picked up by the resource agencies, and they were represented in force at that meeting. But losses of riparian habitat have continued over the intervening years, even as we have learned the true value of these corridors in helping tame the forces at work within the rivers. Central valley riparian forests have been reduced now to barely 1 percent of the original pre-Gold Rush acreage. In many cities and in some heavily grazed areas, the corridors scarcely exist at all. In the valleys these forests are casualties to agricultural and other economic development on the side that borders the uplands. On the side that faces the river they fall prey to limited-purpose water management programs, usually aimed at flood control and delivery of water. The list of benefits from wise management of riparian lands is becoming familiar to people who attend these conferences. Though, as one resource manager put it, it takes a conference like this to remind us that the values are not just those related to the one resource that each of us happens to be concentrating on. The list of riparian values is not endless, but it is long and it includes these: Protects banks from erosion. Helps to reduce the impact of flooding. Provides quality living conditions for fish and wildlife. Creates corridors for their migration. Harbors a number of endangered species. Produces abundant fodder for cattle. Produces timber and other wood products. Provides recreation sites. Contributes to the natural beauty of an area. This conference was convened so that resource managers, researchers, agency administrators, users of the resources, and environmentalists could examine those values, provide an update on their status and management for all who are concerned with this complex of resources, and seek integration of the effort to protect and enhance them. This second big conference had three emphases: (1) improving understanding of the ways that river, channel, bank and living things normally work together as systems in the riparian zone, (2) providing an appreciation for the part that riparian systems play in sustaining populations of several threatened species, and (3) reporting the results of experiments in restoring and revegetating riparian systems. A number of participants have pointed out that this was not actually the second California riparian conference. It was the fourth. David Gaines, who later pioneered the conservationists' effort at Mono Lake, led the way by organizing an initial

conference for about 70 participants in Chico in 1976. This was followed a year later by a similar conference in Davis, organized by Anne Sands. Entitled, "Riparian Forests in California: Their Ecology and Conservation," this memorable conference on the status of the Central Valley riparian forest drew 128 people. Offered in expanded form in 1981, the first "California Riparian Systems Conference" drew 711 people from an incredible array of interests and produced 1035 pages of proceedings still in print, still in demand, and still heavily used. It is, in fact, occasionally used as a textbook. The second California Riparian Systems Conference, which took place on September 22-24, 1988, demonstrated the continued growth of this concern, drawing nearly 900 participants. This was at a time when workshops, training sessions, and focused conferences on riparian habitats had become common. The smaller conferences appear to be serving the training and dissemination functions for a concern that is now well established. Often these smaller meetings have been aimed at specialists in limited fields, e.g., range management, forestry, hydropower or fisheries management. We perceived, therefore, that the big conference should be the place where ideas might be hatched and critiqued, controversies could be aired, and the work of integrating what many of us believe has become too scattered an effort would most definitely be undertaken. The roster of speakers attests to the success the meeting had in drawing together diverse interests. In that list of more than 200 people, agencies loom largest. Surprisingly, consultants were almost as numerous. University contingents were surprisingly large, considering the fact that riparian concerns are largely peripheral to most academic imperatives. The citizens' organizations, resourceoriented private businesses and other user groups were less well represented in the speakers' list, though their presence was felt both in the discussions and in the support that some of them provided in financing the conference. Another mark of the success of the conference is seen in the fact that few participants could agree on what was best about it. For some, a panel on progress in preserving riparian lands along the Sacramento River (which we were not able to reproduce here) was best. Others felt that a panel on integrating public and private interests came closest to meeting their needs. Others were especially satisfied with an evening discussion of ways to interest local communities in preserving stream environments (which also had to be omitted from these proceedings). There were many though who felt that the technical sessions offered the most. It is for people who are likely to share this view that these proceedings are offered, for those are the parts of a conference that can best be reproduced in print. This has to be done, though, while recognizing the fact that the essence of a multifaceted conference like this is really in the spirit that it helps foster. That spirit began with the people who first gave expression to the public concern for riparian lands through meetings like this. Some of the most significant of those people are no longer with us and it is to them that we dedicate this publication, hoping that it will help to continue the movement in a direction and at a pace that would have given them satisfaction. Thus, we dedicate these proceedings to the memory of Richard E. (Rick) Warner, whose vision and devotion to the cause of riparian conservation live on in all of us, and to David Gaines, who started many of us on this journey, and to Mona Myatt, who caught that vision and helped see, through her company's contribution, that this effort could move aheadeven though she could not follow.

ii

This conference was made possible by many people. The sponsors (contributors of $3,000 or more) and co-sponsors (lesser amounts or in-kind contributions) represent a wide range of support and include interests that have often been in conflict. This kind of breadth was seen also in the Advisory Committee, which numbered more than 40 (Appendix), and drew much enthusiastic participation, despite the potential for difference that existed among them. Special thanks are due the Executive Group from that committee: John Menke, JoAnne Sorenson, Ann Riley, Ron Schultze, and Jim Nelson, who contributed much time and were almost never, in the press of their many other duties, heard to say "no" to a request for help. John Stanley, John Rieger, Roland Risser, Deborah Shaw-Warner, Phil Meyer, Steve Chainey and Earle Cummings, were not in the Executive Group but contributed almost as muchalways willingly. The staff of University Extension, with Lynn Read, Audrey Fowler, and Mike McCoy at the top of a long list, helped enormously in preparing for the conference, as did numerous individuals at the Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, in Berkeley, in preparing the Proceedings for publication. Thanks go to 30 people who responded to our need to pass the papers through technical review on schedule that left most of us gasping. Their pleasant and uncomplaining help is gratefully acknowledged. These Proceedings were edited by Bert Schwarzschild and Roberta Burzynski of the Pacific Southwest Station (they also served as Proceedings Editorial Coordinators) and were electronically produced by the Computer Sciences Department of Texas A&M University, College Station, under the direction of Ban Childs. Finally, special thanks go to my wife, Bonnie, who never once complained of my absence and near-total distraction during the months that led up to this conference. Dana L. Abell University of California, Davis Technical Coordinator

iii

SESSION A: CHANNEL DYNAMICS AND RIPARIAN SYSTEMS


Geomorphologists have long been conscious of the need to reckon the bank stabilization effect of riparian vegetation into stream channel equations. Landmark studies were done some years ago by Stanley Schumm on the complex channels of the Murray River in Australia and by M. Gordon Wolman on some streams in the Eastern United States. By and large, though, the processes that control channel geomorphology and the development of riparian vegetation have received entirely separate attention in the years since then. An adequate understanding of the full interactive nature of the channel processes with riparian plant processes has been slow in coming, since a number of factors that are not normally considered in expressions of traditional channel dynamics have turned out to assume significant roles in streams with strong riparian borders. Streamflow dynamics, sediment transport, geology, channel morphology, channel pattern as well as plant form, plant succession and other purely biological processes are all part of the story. In California these factors vary greatly with differing climate, geology and history of human activity. Clearly, the interaction between these two sets of processes changes also under differing patterns of channel alteration and use. An understanding of these changes is going to be essential to planning and design of programs for the protection and rehabilitation of riparian environments in the future. In some of the papers the focus is on the role that riparian trees play in establishing channel boundaries (the Lisle, Barro et al., and Trush et al. papers). Others draw contributions from interdisciplinary teams to link streamside conditions to ecological processes within the stream (the papers from the Oregon State University "stream team", Gregory et al. and Lamberti et al.) or to link human impacts to processes affecting the whole riparian corridor (Buer et al.). The Williams paper and the two Harvey papers deal with the deliberate design of river channels, with and without riparian elements.

Chuck Watson Hydrology Consultant Sacramento, California Dana L. Abell University of California Davis, California

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

INFLUENCE OF VALLEY FLOOR LANDFORMS ON STREAM ECOSYSTEMS1


Stanley V. Gregory, Gary A. Lamberti, and Kelly M. S. Moore Abstract: A hierarchical perspective of relationships between valley floor landforms, riparian plant communities, and aquatic ecosystems has been developed, based on studies of two fifth-order basins in the Cascade Mountains of Oregon. Retention of dissolved nitrogen and leaves were approximately 2-3 times greater in unconstrained reaches than in constrained reaches. Both valley floor landforms and riparian plant communities influenced the abundance of primary producers. Abundances of cutthroat and rainbow trout in unconstrained reaches were approximately twice those observed in constrained valley floors. Valley floors are one of the most physically dynamic components of the landscape, incorporating major agents of terrestrial disturbance and fluvial disturbance. These corridors are major routes for the flux of water, sediments, nutrients, and species. Because of their unique properties, valley floors play an important role in landscape ecology.
2

Most definitions of riparian zones for land management or ecological research are based on a few selected hydrologic, topographic, edaphic, or vegetative attributes of riparian areas. Riparian zones have been investigated from the perspectives of erosion control by riparian vegetation, phreatophyte ecology, uptake of nutrients or contaminants from groundwater, chemistry of water entering lakes and rivers, shading of headwater streams, effects on aquatic invertebrates, migration routes for wildlife, habitat for water fowl, and fish habitat. All of these subjects are critical aspects of riparian ecology, but it is important to recognize the constraints of concepts or definitions of riparian zones developed for specific sets of objectives. In recent decades, ecologists and land use managers have recognized the importance of the structure and functions of riparian zones for both aquatic and terrestrial ecosystems (Knight and Bottorf 1984, Meehan and others 1977, Swanson and others 1982). Meehan and others (1977) defined riparian vegetation as "any extraaquatic vegetation that directly influences the stream environment". From an aquatic perspective, riparian zones are defined functionally as three-dimensional zones of direct interaction with aquatic ecosystems, extending outward from the channel to the limits of flooding and upward into the canopy of streamside vegetation (fig. 1). Examples of critical functions of riparian vegetation for stream ecosystems include shading, bank stabilization, uptake of nutrients, input of leaves and needles, retention of particulate organic matter during high flows, and contribution of large wood.

Riparian zones occur at interfaces ecological interfaces between different ecosystems and conceptual interfaces between different scientific disciplines. The plethora of definitions and delineations of riparian zones are inconsistent, confusing, and often contradictory, reflecting the diversity of disciplines, perspectives, and objectives from which riparian zones have been studied. Most riparian studies have examined selected facets of riparian ecology, but few have developed integrated concepts of the physical, chemical, and biological properties of the interface between aquatic and terrestrial ecosystems. The lack of ecosystem perspectives of riparian areas severely limits our understanding and proper management of these unique components of the landscape. The weakness of ecosystem perspectives in riparian research is evidenced in the term "riparian ecosystems," a term frequently encountered in riparian literature. Riparian zones are interfaces between terrestrial and aquatic ecosystems and exhibit gradients in community structure and ecological processes of these two major ecosystems. Considering riparian zones as distinct ecosystems obscures the patterns of process and structure that are the basis for the great diversity of biota and landforms in riparian areas.

1Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. 2 Associate Professor of Fisheries, Assistant Professor of Fisheries and Graduate Research Assistant, Department of Fisheries and Wildlife,

respectively, Oregon State University, Corvallis, Oregon USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Geomorphology of River Valleys


Flowing water interacts with the parent geology and inputs of organic and inorganic material from adjacent vegetation and hillslopes to form channels and floodplains within river valleys. Most geomorphic research has focused on lowland, floodplain rivers. In these alluvial systems, channel migration creates valley floor landforms and surfaces for development of riparian plant communities. The geomorphic dynamics of riparian zones in steep mountain landscapes involve both fluvial processes and mass movement events of adjacent hillslopes. Valley floors contain both active channels and adjacent floodplains and terraces. Active channels are separated from adjacent hillslopes or floodplains by distinct topograhic breaks and commonly represent the lower extension of perennial terrestrial vegetation (Redman and Osterkamp 1982). Floodplains are relatively flat surfaces that are formed by fluvial deposition of sediments adjacent to active channels and are inundated during major floods. Several floodplain surfaces may occur within a valley floor; successively higher surfaces are flooded at less frequent intervals. In lowland valleys, floodplains are submerged much more frequently and for longer duration than floodplains in mountainous terrain. All floodplains and active channels are bordered by the lower flanks of adjacent hillslopes. A drainage network extends from the headwaters to estuaries. Sections of a drainage network are differentiated by major topographic discontinuities, such as high gradient montane rivers, low gradient, lowland rivers in broad valleys, and broad coastal plains. Segments are continuous areas within a drainage formed by common large-scale geomorphic processes, and they have different potentials for development of active channels and floodplains. A drainage segment is composed of reach types, delineated by the type and degree of local constraint imposed by the valley wall at the channel margin. The degree of local constraint controls the fluvial development of geomorphic surfaces. Constrained reaches (valley floor narrower than two active channel widths) occur where the valley floor is constricted by bedrock, landslides, alluvial fans, or other geologic features. Streams within constrained reaches are relatively straight, single channels with little lateral heterogeneity. River valleys in constrained reaches are narrow and include few floodplains. Consequently, riparian vegetation in these areas is similar in composition to adjacent hillslope plant communities. Unconstrained reaches (valley floors wider than two active channel widths) are less constrained laterally
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 Riparian zone defined in terms of zones of influence of streamside vegetation on stream ecosystems (Meehan and others 1977).

Shading by streamside vegetation influences water temperature and aquatic primary production. Rooting of terrestrial vegetation within and adjacent to stream channels stabilizes banks and minimizes soil erosion. Living vegetation provides complex habitats for both aquatic and terrestrial wildlife. Leaves and needles provide essential food resources and habitat for aquatic insects. When floodplains are inundated, stems and roots of floodplain trees and shrubs comb organic matter out of transport and store it for subsequent processing on the floodplain or in the stream. Woody debris provides important structural elements that protect and stabilize stream banks, stores sediments that serve as habitat and spawning gravel, and traps organic matter that provides food for aquatic organisms. Dead woody material, both standing snags and wood on the forest floor, provides essential habitat for wildlife.

Most often management agencies adopt operational definitions of riparian zones that are based on hydric soils and unique terrestrial plant associations. If management agencies adopt perspectives of riparian zones that do not address critical ecosystem processes, the integrity of riparian resources cannot be insured. Integration of aquatic biological processes with physical templates of channel geomorphology and hydraulics has been a major challenge for stream ecologists in recent years. We present a hierarchical perspective of relationships between valley floor landforms, riparian plant communities, and aquatic ecosystemsan integrated ecosystem perspective of riparian zones. 4

and provide greater potential for floodplain development and active channel migration. Unconstrained reach types exhibit complex, braided channels and extensive floodplains, which support a diverse array of plant communities of different age. Riparian stands in these areas include many plant species adapted to frequent flooding. Reach types are made up of sequences of channel units, representing different bed forming processes. Channel units in low gradient, sand and gravel bed streams are generally classified simply as pools and riffles (Leopold and others 1964). In high-gradient streams with coarser bed material, the distinction between high and low gradient units is conspicuous, but the steeper units may be divided into several additional types rapids, cascades, falls. With the exception of abrupt falls, channel units are longer than one channel width and are distinguished on the basis of surface slope, degree of turbulence, and extent of supercritical flow. Channel sub-units include hydraulic and geomorphic features shorter than the active channel width. Rifles, pools, rapids, and other features that are shorter than one channel width are categorized as sub-units. Backwaters, eddies, and side channels are also sub-units and play a distinctly different ecological role than subunits along the main axis of the channel. Sub-unit features correspond to the habitat types employed in most aquatic ecological research. As flow increases and the active channel is completely inundated, channel units attain uniform surfaces and delineations between subunits are obscured.

richness between hillslope and riparian areas were found in plant communities of the Sierra Nevada in California (table 1). Patterns of disturbance, particularly flooding, influence the spatial complexity of riparian plant communities, while environmental gradients, such as light, temperature, and moisture, determine the sharpness of transitions between riparian and hillslope plant communities.
Table 1 - Number of plant species per plot in riparian zones of the Cascade Mountains of Oregon and the Sierra Nevada of California (Art McKee, personal communication). Number of Plant Species Upland Riparian 19-32 28-38 51-107 51-55

Location Cascade Mountains McKenzie River Sierra Nevada Sequoia National Park

Riparian Vegetation
Riparian zones are extremely patchy, reflecting past fluvial disturbances from floods and non-fluvial disturbances of adjacent hillslopes. Fluvial processes are the dominant disturbance in riparian areas, but fire, wind, landslides, drought, freezing, disease, insects, and other natural agents of mortality common on upper slopes may influence riparian stands along valley floor. Furthermore, topoedaphic conditions of valley floors are extremely varied, ranging from continuously wet to extremely dry over short distances. Consequently, riparian plant communities are structurally and taxonomically diverse. In conifer forests of the northwest, riparian plant communities exhibit greater diversity than plant communities of upper hillslopes. In riparian plant communities ranging from recent clearcuts to old-growth forests in excess of 500 years in Oregon, riparian communities contained approximately twice the species richness observed in upslope communities. Similar contrasts in species

Patterns of disturbance in riparian zones differ from disturbances on upper hillslope in shape and areal extent. Flooding in river valleys creates narrow, linear disturbance patches. The resulting floodplain forest is composed of thin bands of early seral stages, predominantly deciduous. Longitudinally, patches of riparian plant communities commonly are short and alternate from one side of the channel to the other. Within a flood event, the total area disturbed may exceed tens to hundreds of square kilometers, though the width of the disturbance at any point may be only a few tens to hundreds of meters. On a basin scale, the total area disturbed in a given flood may equal or exceed that of common terrestrial disturbances such as fire, wind, and disease. However, individual patches within the disturbed area in floods are small, relative to the overall disturbance, and extremely numerous. The heterogeneity imposed by flooding in river valleys contributes to the biological diversity of riparian areas.

Aquatic Ecosystems
Physical processes within riparian zones influence the biological organization and rates of processes of stream ecosystems. Active channels and floodplains form the physical habitat for aquatic organisms. Large organic matter contributed by riparian vegetation serves as a dominant geomorphic element, particularly in headwater streams (Swanson and Lienkaemper 1978). Riparian vegetation supplies organic matter in the form of leaves, needles, and wood to streams and floodplains. This terrestrial source of organic matter provides a major portion of the food base for stream ecosystems (Cummins 1974). Leaves of herbs, shrubs,

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

and deciduous trees have higher food value for most aquatic invertebrates than the more refractory needles of conifers (Triska and others 1982). Diverse riparian plant communities in broad floodplain reaches potentially offer higher quality food than conifer-dominated riparian zones, but the input of litter from deciduous plants is restricted to a shorter time interval in autumn. Thus, mature conifer forests provide a more consistent but lower quality food supply to stream ecosystems (Gregory and others 1987). The canopy of streamside forests potentially shades the stream channel, decreasing solar radiation available for aquatic primary production. In small, headwater streams, riparian canopies strongly limit primary production, and as streams widen downstream, the influence of riparian vegetation on primary production decreases as the canopy opens over the channel. In this sense, the presence of riparian vegetation reduces aquatic productivity through the algal food base. Removal of riparian vegetation by man also increases solar radiation reaching headwater streams and potentially increases primary production. In Lookout Creek in the McKenzie River drainage, percent cover of filamentous algae was 3-30 times greater in a reach flowing through young second-growth riparian stand than a reach flowing through a 450-year-old old-growth stand. The influence of riparian canopy cover on aquatic primary production is most pronounced in headwater streams and diminishes downstream as the opening over the stream increases with increasing channel width. As a result, the effects of riparian timber harvest on aquatic primary production is relatively greater in headwater streams and decreases downstream. Algal food resources for aquatic organisms are much less abundant in streams than terrestrial litter but much higher in quality as food for invertebrates. Food resources, whether aquatic or terrestrial in origin, must be retained in the stream before being consumed by aquatic organisms. Valley landforms and adjacent riparian plant communities directly influence bed form and channel roughness, which determine retention of water and both dissolved and particulate inputs during both low flow and floods. In two fifth-order basins in the Cascade Mountains of Oregon, we measured the retention of leaves in constrained and unconstrained reach types (reported by Lamberti and others in these proceedings). Leaves in transport in unconstrained reaches with broad floodplains were retained 4-5 times more efficiently than leaves in constrained reaches. Large logs and smaller branches and twigs supplied by riparian vegetation form complex accumulations, which increase the retention efficiencies of stream reaches. In streams of the Cascade Mountains, an average leaf traveled more than 12 m in reaches influenced by debris dams; but an average leaf traveled less than 5 m in reaches influenced by debris accumulations and less than

1 m in reaches that were completely obstructed by debris dams (Speaker and others 1984). Riparian zones modify the cycling of dissolved nutrients as they are transported from hillslopes, across floodplains, and down drainages. In coniferous and deciduous riparian zones of Oregon, microbial transformation of nitrogen were greater in riparian areas than in upper hillslopes (Mike McClellan, Oregon State University, unpublished data). Rates of denitrification were more than five times greater in floodplains than adjacent hillslopes (table 2), and rates of denitrification were higher in alder stands than in coniferous forests. Because of the rapid cycling of nitrogen in the riparian zone, elevated concentrations of nitrate were not observed in streams in alder stands, even though nitrogen fixation was observed.
Table 2 Rates of denitrification in riparian zones and upslopes in a 40-year-old deciduous and a 450-year-old riparian forest (expressed as ng N/g dry weight of soil/hr with standard errors in parentheses). Geomorphic Surface Soil Depth Floodplain Toeslope 0-15 cm 0-30 cm 6.3 (2.2) 0.4 (0.3) 4.2 (4.2) 0.2 (0.1) 8.2 (4.3) 1.4 (1.0)

Site Coniferous Deciduous

Hillslope 1.2 (1.2) 0 3.0 (1.2) 1.2 (0.7)

0-15 cm 15.0 (2.9) 15-30 cm 11.3 (2.6)

Nutrient outputs from watersheds are not only modified within floodplain soils, but nutrients are rapidly taken up and transformed by stream communities as well. In streams of the McKenzie River drainage, we measured uptake of dissolved ammonium in constrained and unconstrained reaches. Dissolved nitrogen (ammonium) was approximately 2-3 times greater in unconstrained reaches than in constrained reaches (reported by Lamberti and others in these proceedings). Unconstrained valley floors are more complex environments both geomorphically and hydraulically and retain water and dissolved nutrients longer, increasing the potential for biological uptake. In addition, unconstrained reaches may support more abundant algal assemblages, increasing the biological demand for nutrients. Uptake of ammonium in reaches of Lookout Creek in young secondgrowth riparian forests was more than twice the uptake observed in reaches flowing through old-growth forests (reported by Lamberti and others in these proceedings), reflecting the influence of primary producers on nutrient cycling. Higher trophic levels are also influenced by valley landforms. In the two study drainages in the McKenzie River in Oregon, abundances of cutthroat and rainbow trout in unconstrained reaches (120-200 individuals/100 m) were approximately twice those observed in constrained valley floors (60-80 individuals/100 m) (reported by Moore and Gregory in these proceedings).

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Unconstrained stream reaches contain broad floodplains with numerous eddies, backwaters, and side channels. In addition to the main channel, these complex channel forms create a greater diversity of fish habitats and provide numerous lateral refuges during floods. In contrast, constrained reaches offer few refuges in which to escape the torrential flows of winter floods. Unconstrained reaches in our study streams also contained greater numbers of trout fry than constrained reaches. Salmonid fry rear in shallow, low velocity habitats along the edges of streams and in side channels and backwaters, particularly those associated with complex floodplains (reported by Moore and Gregory in these proceedings). Thus, the complexity of broad floodplains is beneficial for rearing of new year classes of fish and survival for fish of all age classes.

components of the landscape. In addition, riparian areas are major routes for the flux of water, sediments, nutrients, and plant and animals within drainage networks. Because of their unique properties, riparian areas play important roles in landscape ecology and resource management.

Acknowledgements
We thank Linda Ashkenas, Randy Wildman, and Al Steinman for their assistance in data collection and analysis and Fred Swanson and Gordon Grant for their assistance in conceptual development and for providing unpublished geomorphic data. This research was supported by grant number BSR-8508356 from the National Science Foundation.

Conclusions References
From an ecosystem perspective, riparian areas are created and maintained through extensive interaction between valley landforms, succession of terrestrial vegetation, and the structural and functional responses of aquatic ecosystems. Geomorphic processes create the structure of stream channels and floodplains, which serve as physical templates for successional development of riparian vegetation. The structure and function of stream ecosystems are strongly influenced by the habitat and food resources provided by channel structure and streamside vegetation. Resource management agencies are faced with the pragmatic problem of identifying boundaries on landscapes without abrupt demarcation. Although effective riparian management requires establishment of such riparian management zones, all managers must constantly remind themselves that their riparian management zones usually include only a portion of the interface between terrestrial and aquatic ecosystems. Recognition of the trade-offs inherent in any riparian management system requires ecologically robust concepts of riparian areas. Management concepts and definitions of riparian areas that exclude the physical, chemical, and biological interactions within the interface between terrestrial and aquatic ecosystem cannot insure the ecological integrity of one of the most physically dynamic components of the landscape. The riparian areas along river valleys experience many of the disturbances of upslope forests (e.g., fire, disease, insect outbreak, wind) as well as the unique disturbance associated with floods. Riparian areas are also interfaces between terrestrial and aquatic ecosystems, encompassing overlapping gradients in the physical and biological properties of these distinctly different ecosystems. As a result, riparian areas are one of the most physically complex and biologically diverse
Cummins, K.W. 1974. Stream ecosystem structure and function. BioScience 24:631-641. Gregory, S.V.; Lamberti, G.A.; Erman, D.C.; Koski, K.V.; Murphy, M.L.; Sedell, J.R. 1987. Influence of forest practices on aquatic production. In: Salo, E.0; Cundy, T.W., eds. Streamside Management: Forestry and Fishery Interactions Symposium; February 12-14, 1986, Seattle, WA: Institute of Forest Resources, University of Washington; 234-255. Hedman, E.R.; Osterkamp, W.R. 1982. Streamflow characteristics related to channel geometry of streams in western United States. Water Supply Paper 2193. Alexandria, VA: U.S. Geological Survey; 17 p. Knight, A.W.; Bottorff, R.L. 1984. The importance of riparian vegetation to stream ecosystems. In: Warner, R.E., Hendrix, K.M., eds. Proceedings, California Riparian Systems, Berkeley, CA: University of California Press; 160167. Lamberti, G.A.; Gregory, S.V.; Ashkenas, L.R.; Wildman, R.C.; Steinman, A.D. 1989. Influence of channel geomorphology and riparian zones on nutrient retention in stream ecosystems. (These proceedings). Leopold, L.B.; Wolman, M.G.; Miller, J.P. 1964. Fluvial processes in geomorphology. San Francisco, CA: W.H. Freeman; 522 p. McKee, W.A., Department of Forest Sciences, Oregon State University, Corvallis, OR. (Unpublished data provided by author). 1988. Meehan, W.R.; Swanson, F.J.; Sedell, J.R. 1977. Influences of riparian vegetation on aquatic ecosystems with particular references to salmonid fishes and their food supply. In: Johnson, R.R.; Jones, D.A., eds. Symposium, Importance, preservation and management of riparian habitat; USDA For. Serv. Gen. Tech. Rep. RM-43. Fort Collins,

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

CO: Rocky Mountain For. and Range Exp. Stn., Forest Service, U.S. Dept. of Agriculture; 137-145. Moore, K.M.; Gregory, S.V. 1989. Geomorphic and riparian influences on the distribution and abundance of salmonids in a Cascade Mountain stream. (These proceedings). Speaker, R.W.; Moore, K.M.; Gregory, S.V. 1984. Analysis of the process of retention of organic matter in stream ecosystems. Verh. Internat. Verein. Limnol. 22:18351841. Swanson, F.J.; Lienkaemper, G.W. 1978. Physical consequences of large organic debris in Pacific Northwest s t r e a m s . U S D A F or . S e r v . Ge n . Te c h . R e p. P NW 6 9. Portland, OR: Pac. Northwest For. and Range Exp.

Stn., Forest Service, U.S. Dept. of Agriculture; 12 p. Swanson, F.J.; Gregory, S.V.; Sedell, J.R.; Campbell, A.G. 1982. Land-water interactions: the riparian zone; In: Edmonds, Robert L., ed. Analysis of Coniferous Forest Ecosystems in the Western United States. US/IBP Synthesis Series 14. Hutchinson Ross Publishing Co., Stroudsburg, PA; 267-291. Triska, F.J.; Sedell, J.R.; Gregory, S.V. 1982. Coniferous forest streams. In: Edmonds, Robert L., ed. Analysis of Coniferous Forest Ecosystems in the Western United States. US/IBP Synthesis Series 14. Hutchinson Ross Publishing Co., Stroudsburg, PA; 292-332.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

CHANNEL-DYNAMIC CONTROL ON THE ESTABLISHMENT OF RIPARIAN TREES AFTER LARGE FLOODS IN NORTHWESTERN CALIFORNIA 1
Thomas E. Lisle2
Abstract: Large floods in northwestern California in the past two decades have mobilized extensive areas of valley floors, removed streamside trees, and widened channels. Channel cross sections were surveyed to illustrate an hypothesis on the linkage between sediment transport, colonization of channel margins by trees, and streambank recovery. Riparian trees, e.g., white alder (Alnus rhombifolia), colonize the water's edge at low flow to receive adequate moisture during the dry season. Such stands can endure annual high flows only after the flood-enhanced sediment load declines and the width of the annually mobile bed contracts to the low-flow width. Streambank formation along the low-flow margin can then proceed by deposition of fine sediment and organic debris. A series of large floods from 1950 to 1975 (Harden and others 1978) greatly altered riparian ecosystems in north coastal California from the Eel River basin northward. Channel aggradation as much as several meters, channel widening as much as 100 percent, and extensive destruction of trees by flood waters widened the zone of active bed sediments at the expense of riparian corridors along many streams (Hickey 1969; Kelsey 1980; Lisle 1981). Although most aggraded channels have degraded to stable bed elevations as excess sediment has been transported downstream, many remain widened (Lisle 1981). If probability prevails and such large floods do not recur, how will riparian stands and associated streambanks recover over the next few decades? This paper presents a hypothesis on a relation between colonization of streamside trees and channel dynamics that may govern the recovery of riparian stands and reconstruction of streambanks. During large floods, extensive areas of streambeds and floodplains can be mobilized by high shear stresses and new inputs of sediment. Riparian species of willow and alder that have low tolerance to moisture stress tend to colonize the water's edge during summer low-flow periods. Because these trees also require a stable substrate, they can establish only after the zone of annually mobilized bed material contracts to a small fraction of the width mobilized by the last large flood. Once established, the trees can trap fine sediment and organic debris, add root strength to bed material, reduce local shear stress, and thereby induce formation of new streambanks. Widened channels thus may recover when new streambanks form inside flood-eroded banks at a spacing dictated by the zone of annually mobilized bed sediment.

Trees Along Mobile Bed Margins


Channel cross sections showing substrate and vegetation of three creeks affected by recent floods (1964, 1972, and 1975) in north coastal California illustrate colonization and growth of riparian trees along mobile bed margins. All examples were surveyed across reaches where bed and banks were composed of alluvium. The first example (Prairie Creek) presents the hypothesis in detail and illustrates bank formation along a channel transporting abundant fine sediment. The second (Hurdygurdy Creek) shows contrasts with a channel transporting little fines. The third (Willow Creek) details plant species occupying micro-habitats within the channel.

Prairie Creek Prairie Creek, a tributary of Redwood Creek, Humboldt County, with a drainage area (DA) of 34 km2 at the study reach, has a moderate sediment yield with abundant fines (Lisle, in press). The basin is heavily forested mostly with old-growth redwood. Summers are relatively cool and moist because the basin is only a few kilometers from the coast. Tertiary sands and gravels of the Gold Bluffs Formation (Moore and Silver 1968) underlying much of the basin supply readily mobile bed material. Flood flows usually cause modest changes in channel morphology, however, because human disturbance has been relatively light, the channel gradient is low (0.0032 in the study reach), and streambanks are strengthened by dense riparian vegetation. The mobile portion of the streambedthat which is entrained annuallyis composed of sand and gravel with median grain diameter (D50) of 9.0 mm and is armored with pebbles and cobbles (D50=25.5 mm) (Lisle, in press).

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Research Hydrologist, Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, Arcata, Calif.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 Cross section of Prairie Creek, showing bed elevation changes from 1980 to 1988 and the distribution of riparian vegetation and substrate. Approximate bankfull elevation is indicated by BF. A natural levee approximately 0.7 m high had accreted progressively on the inner right bank immediately adjacent to the mobile bed from 1980 to 1988 (fig. 1). The levee, which was composed of fine sand, silt, and organic debris, supported a narrow dense stand of red alder (Alnus oregona). The levee was highly irregular, due to locally thick deposits of newly laid fine sediment and local scour around large woody debris caught among the alders. The alders were rooted at about one-half of bankfull stage, and thus were probably flooded several times on average each year. Bankfull elevation in this cross section was poorly defined but, by extrapolation from adjacent reaches, corresponds approximately with the top of the bar. In contrast to the levee, the adjacent bar surface was smooth and covered with low herbaceous vegetation. The bar accreted less than 0.3 m from 1980 to 1988. The bar surface in 1980 was composed of sand and pebbles and scant herbaceous vegetation; by June 1988, a denser mat of vegetation and organic matter had accumulated. This mat was probably more efficient at trapping fine sediment than the sparsely vegetated surface of 1980. Repeated surveys of this cross section since 1980 suggest the following sequence of processes in forming the levee on the inner right bank. A major flood occurred in the Redwood Creek basin in 1975 (Harden and others 1978). Tree-ring dating of the largest trees on the levee indicate that trees within the channel were stripped by the flood but reestablished within a few years. The entire streambed was extensively reworked by the flood, or in other words, the bed apparently became mobile over its entire width (39 m). Afterward, the mobile zone of the bed contracted to approximately its width at the time of study (10m), and alder were able to colonize the low-water's edge along the left margin of the bar. At this time the bar surface merged smoothly with the mobile bed, as indicated by the 1980 survey. During even modest stormflows, high

friction generated by alders rooted low in the water created steep velocity gradients away from the main flow and promoted rapid deposition of fine sand, which is only intermittently suspended. Sand comprises as much as 55 percent of suspended sediment in Prairie Creek (Lisle in press). Sedimentation rates over the remainder of the bar were slower because of its smoothness and the depletion of coarse suspended fractions at the levee. As a result, the levee originated at a lower elevation but built faster than the bar, so that a level floodplain may be formed eventually. Channel incision since 1980 apparently helped to define a higher bank as the levee built (fig. 1). Root strength probably helped to stabilize these higher banks.

Hurdygurdy Creek

Hurdygurdy Creek (DA = 70 km2 at the study reach), a major tributary of the South Fork Smith River, Del Norte County, is coarser and steeper than Prairie Creek and carries little fine sediment. In the study reach, the bed surface is mainly cobbles and boulders (D50 = 155 mm), and channel slope is 0.02. Moisture stress on riparian vegetation is greater than along Prairie Creek, because Hurdygurdy Creek is farther from the coast (20 km) and summers are hotter and drier. Sediment is pre-dominantly coarse bedload derived from metamorphic rocks of the Mesozoic Age. Unpublished data provided by Mike Furniss, Six Rivers National Forest, Eureka, California shows that the channel appears stable and cross sections surveyed since 1976 show little change. White alders (Alnus rhombifolia) grew within the channel of Hurdygurdy Creek, but in contrast with trees in Prairie Creek, were not associated with bank formation (fig. 2). The alders were rooted near the lowflow water surface elevation, which was at less than one-half of bankfull stage. At this cross section, alders on the left portion of the channel ended abruptly at a certain point that was not related to elevation, but instead probably defined the left margin of the mobile bed. The bed to the right of the alders had an equal elevation and thus equal availability of moisture. This area was an active bar deposited against the larger, stabilized flood bar upon which the alders had established. The bed among the alders was clearly not mobile enough to cause removal of the trees. The alders grew among mossy boulders that had not moved apparently since the flood of 1964. Cores extracted from the largest trees indicated colonization of the bar was no later than 1966. The alders had trapped abundant large woody debris but little fine sediment.

10

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

As in Prairie Creek, riparian trees were rooted densely at the low-flow water's edge and appeared to stabilize alluvium sloping toward the channel thalweg. The trees thereby helped to define streambanks that were well below bankfull stage. As in Hurdygurdy Creek, however, little fine sediment deposited among the trees, which would lead to bank formation.

Figure 2 Cross section of Hurdygurdy Creek surveyed in July 1986, showing growth pattern of riparian trees and substrate. BF indicates bankfull elevation. Willow Creek Willow Creek (DA = 110 km 2 at the study reach) is similar to Hurdygurdy Creek in climate and geology. The study site is 0.5 km upstream of the Highway 96 bridge and 1 km upstream of the junction of Willow Creek with the Trinity River in Humboldt County. Cartographic channel slope in the study reach is 0.019. The 1964 flood destroyed the stream gauge, which was located in the study reach, and deposited sediment on the prominent terrace (T2, fig. 3) 4-5 m above the existing streambed. Unrecorded large floods occurred probably in 1972 and 1975, as well; high rainfall and runoff were recorded in Redwood Creek, the adjacent basin to the west (Harden and others 1978). Wet-site trees and herbs, including white alder, willow (Salix hindsianai; S. lasiolepis), black cottonwood (Populus trichocarpa), and sweet clover (Melilotus albus), grew in a narrow band (2-5 m wide) along the low-flow water's edge (Zones 1 and 2 in Section 1, Zone 1 in Section 2, fig. 3; table 1). Riparian alders and willows were established no higher than about 0.4 m above a stage typical of late summer, which is approximately 0.3 m below that during the survey. Bankfull was poorly defined, but corresponded roughly with the top of the bar in Section 1. Riparian trees were rooted at less than one-half of bankfull stage as defined. Cores extracted from the largest trees indicated colonization in 1973-1975. Herbaceous species growing on higher, drier sites (Zones 3 and 4) were typical of those that invade disturbed ground and have a wide range of moisture tolerances. The base of the drier zones was 0.2 m (Section 1) and 0.9 m (Section 2) above late summer stage.

Figure 3 Partial cross sections of Willow Creek surveyed in June, 1988, showing distribution of riparian vegetation and substrate. Numbered brackets show vegetation zones of table 1. BF indicates approximate bankfull elevation; T1 and T2 are alluvial terraces.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

11

Table 1 Species, relative abundance, and indicated habitat conditions of vegetation along zones of Sections 1 and 2 (fig. 3), Willow Creek.

Species1 and abundance2 Section 1: Zone 1 Salix hindsiana-A (sandbar willow), Populus trichocarpa-S (black cottonwood), Mellilotus albus-C (sweet clover) Salix hindsiana-S, Mellilotus albus-S

Indicated habitat

High year-round moisture High moisture

Zone 2

Zone 3

Mellilotus albus, Silene gallica (catchfly), Bromus sp., Erodium cicutarium (filaree), Aira Caryophylla (hair grass), Kohlraushia velutina, Medicago polymorpha (bur-clover), Torilis arvensis, Vicia americana (vetch), Trifolum sp.3

Recent or frequent disturbance; variable moisture requirements

Section 2: Zone 1 Alnus rhombifolia-A (white alder), Rumus crispus-S (curly dock), Mellilotus albus-C unknown grass-C Mellilotus albus-A, Kohlrauschia velutina-C, Silene gallica-S, Lotus sp.-S, Lupinus sp.-S Silene gallica, Bromus sp., Aira caryophyllea, Taraxacum officinale (dandelion) Cirsium sp., Trifolium sp., Lotus sp 3 Bromus sp. (and other grasses)-A, others:3 Lotus sp., Trifolium sp., Vicia americana, Cirsium sp., Salix lasiolepis (arroyo willow) High moisture

Zone 2

Disturbance

Zone 3

Disturbance

Zone 4

Stabler substrate than Zone 3

Scientific names follow Munz (1968). A-abundant, C-common, S-sparse All species approximately equal in abundance: sparse to common.

12

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Discussion and Conclusions


One might expect flood-widened channels to become narrower by accreting material against eroded streambanks. Post-flood surveys of cross sections at gauging stations, however, have shown this not to be the case (Lisle 1981). Instead, observations of incipient bank formation and growth patterns of riparian trees suggest that recovered stream margins may become defined well within flood-eroded channels and below bankfull stage, and then accrete vertically. The first step in this process is for sediment load to decrease to a level whereby the zone of annually mobile bed material contracts and becomes confined to a zone corresponding to summer flow margins. Riparian trees, especially white alder which require both a stable substrate and a shallow rooting depth to year-round moisture (Griffin and Critchfield 1976), can then establish along low-flow margins, even though they are frequently inundated during the wet season. With further growth, trees can significantly stabilize bank materials by adding root strength and reducing local shear stress through added roughness. In streams such as Prairie Creek that carry abundant fine sediment, trees may also promote deposition of suspended sediment, particularly sand and organic debris, and thereby cause banks to build vertically. In streams such as Hurdygurdy and Willow Creeks that are steeper and carry little fine sediment, banks may not readily form but instead become defined by channel incision between densely rooted riparian corridors. This mechanism for the recovery of channels and riparian stands could apply to a variety of gravel-bed streams in Mediterranean climates. Observations of stages in colonization and growth of riparian trees and their substrate could indicate much about the condition and recent geomorphic history of the channel. The presence of narrowly spaced riparian corridors inside flood channels indicates that sediment load has waned considerably since the last flood, and that the channel has an extensive width to accommodate an increased load. The absence of trees within bankfull margins, on the other hand, may indicate that the entire width of the bed was mobilized recently and that an increased load may lead to further bank erosion or bed aggradation. The hypothesis has not been tested. It suggests many opportunities, however, for fruitful collaborations between plant ecologists and fluvial geomorphologists seeking to understand channel recovery from floods.

Acknowledgments
I thank Bruce Bingham for identifying species at Willow Creek, researching their habitat requirements, helping to survey the cross sections, and reviewing the manuscript; and Lori Dengler for reviewing the manuscript.

References

Griffin, James R.; Critchfield, William B. 1976. The distribution of forest trees in California. Research Pub. PSW82/1972. Washington, DC; U.S. Department of Agriculture, Forest Service; [reprinted with supplement, 1976] 118 p. Harden, Deborah R.; Janda, Richard J.; Nolan, K. Michael. 1978. Mass movement and storms in the drainage basin of Redwood Creek, Humboldt County, Californiaa progress report. U.S. Geological Survey Open-File Rep. 78-486, 161 p. Hickey, John J. 1969. Variations in low-water streambed elevations at selected stream-gauging stations in Northwestern California. U.S. Geological Survey Water-Supply Paper 1879-E, 33 p. Kelsey, Harvey M. 1980. A sediment budget and an analysis of geomorphic process in the Van Duzen River basin, north coastal California, 1941-1975. Geological Society of America Bulletin 91(4): 1119-1216. Lisle, Thomas E. 1981. Recovery of aggraded stream channels at gauging stations in northern California and southern Oregon. In: Davies, Timothy R.H.; Pearce, Andrew J., editors. Proceedings, Erosion and Sediment Transport in Pacific Rim Steeplands. IAHS-AISH Pub. 132: 189-211. Lisle, Thomas E. Deposition of fine sediment in natural gravel streambeds. Water Resources Research [in press]. Moore, George W.; Silver, Edward A. 1968. Geology of the Klamath River delta, California. U.S. Geological Survey Prof. Paper 600-C; C144-C148. Munz, Phillip A. 1968. A California Flora with Supplement. University of California Press, 1681 p.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

13

ALDER ESTABLISHMENT AND CHANNEL DYNAMICS IN A TRIBUTARY OF THE SOUTH FORK EEL RIVER, MENDOCINO COUNTY, CALIFORNIA 1
William J. Trush, Edward C. Connor and Allen VV. Knight Abstract: Riparian communities established along Elder Creek, a tributary of the upper South Fork Eel River, are bounded by two frequencies of periodic flooding. The upper limit for the riparian zone occurs at bankfull stage. The lower riparian limit is associated with a more frequent stage height, called the active channel, having an exceedance probability of 11 percent on a daily average flow duration curve. Distinct tree communities occupy bankfull and active channel zones. Riparian densities (trees per meter of stream channel) along the active channel decreased with increasing channel gradient and curvature. Riparian densities at bankfull stage were not as sensitive to change in channel gradient and curvature.
2

Bankfull and active channels are discernable alluvial features on Elder Creek, a bedrock tributary to the South Fork Eel River. Elder Creek is the largest pristine Douglas-fir watershed in Northern California. The watershed is managed jointly by the Nature Conservancy and Bureau of Land Management, U.S. Department of the Interior, as the Northern California Coast Range Preserve. In the main channel, the lower limit for permanent woody vegetation ranges from the active channel crest up to bankfull stage, depending on location within the channel. This paper examines the relationships between riparian tree populations and stream morphology in a pristine, bedrock channel. Specifically, the following hypotheses are examined: 1. Streambank species compositions are different at bankfull and active stages; 2. Riparian tree density at active channel stage decreases with increasing channel gradient and curvature; 3. Channel bend inflections support a disproportionately large percentage of the total riparian community; 4. Inside banks support higher riparian densities than the outside banks of channel bends.

Many alluvial streams have distinct banks that overflow only during periods of flood discharge. The flood discharge that just reaches the crest of the bank has been labeled bankfull discharge (Richards 1982; Wolman and Leopold 1957). The frequency of bankfull discharge for many alluvial channels ranges from 1.5 to 3.0 years (or greater) on an annual maximum flood series (for review, Williams 1978). On the tributaries of the upper South Fork Eel River, in Mendocino County, California, streambanks are composed of bedrock or coarse alluvium and do not exhibit the sharp break in channel cross section common in alluvial streams. Rather, bankfull stage height (water surface elevation above the streambed at bankfull discharge) often corresponds to the upper limit of sand deposition or to the upper border of large point bars (Trush, Connor and Knight 1988). A distinct break in channel cross section is apparent at a sub-bankfull stage, forming a smaller channel within the bankfull channel. Osterkamp and Hedman (1977) call this inner region the "active channel." They describe the active channel for Virginia rivers as:
A short term geomorphic feature subject to change by prevailing discharges. The upper limit is defined by a break in the relatively steep bank slope of the active channel to a more gently sloping surface beyond the channel edge. The break in slope normally coincides with the lower limit of permanent vegetation so that the two features, individually or in combination, define the active channel reference level. The section beneath the reference level is that portion of the stream entrenchment in which the channel is actively, if not totally, sculptured by the normal process of water and sediment discharge.

Study Site

Elder Creek watershed (16 km2) is located in the upper South Fork Eel River basin (fig. 1) on the western flank of Cahto Peak (1290 meters). In this region, landslides are a dominant geomorphic process in the highly fractured Franciscan sedimentary landscape (James 1983). More than 95 percent of the high annual rainfall (2030 mm) falls from November through April. Eighty five percent of the annual precipitation can occur as runoff (Rantz 1964). At 0.60 km from the mouth, the U.S. Geological Survey maintains a hydrologic benchmark station (Sta. No. 11475560), monitoring stream discharge and precipitation at 15-minute intervals.

1 Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.
2

Graduate Student, Department of Forestry and Natural Resources, University of California, Berkeley; Graduate Student, and Professor, Department of Land, Air, and Water Resources, University of California, Davis. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

14

(1975) stream classification system Elder Creek would be a 'B1' stream type. Multiple terrace sets 3 to 15 m above the present channel indicate an extensive period of downcutting; significant floodplain formation is limited to very few, less confined channel reaches. Stream gradient for the entire main channel (5.1 km) averages 3.3 percent, though the channel gradient below a major knickpoint (2.3 km from the mouth) is 2.4 percent. The low sinuosity channel does exhibit depositional features typical of alluvial channels, though modified by bedrock outcrops and boulders. Coarse point bars are found at more acute channel bends associated with the larger pools. Pools and riffles have similar median particle size distributions ranging from 15 to 50 cm. Large woody debris has little impact on channel morphology in this pristine watershed.

Active and Bankfull Channels

Many of the criteria for identifying bankfull stage in alluvial channels (Williams 1978) apply to the active channel in Elder Creek. Kush, Connor and Knight (1988) developed the following criteria for identifying the active channel: a. Base of alder trunks occur at active channel crest along straight channel reaches and at bend inflections; on acute channel bends, alders generally occur closer to bankfull stage; b. Root wads of living sedges rarely above active stage unless there is seepage from the banks; Figure 1 Location map of Elder Creek in the upper South Fork Eel watershed. The South Fork Eel River flows due north. Upper watershed slopes are primarily a mix of Ceanothus spp., madrone (Arbutus menziesii), and manzanita (Arctostaphylos spp.). Middle and lower slopes are dominated by Douglas-fir (Pseudotsuga menziesii), tanoak (Lithocarpus densiflorus), and California bay laurel (Umbellularia californica). The riparian tree community is comprised of four major species. White alder (Alnus rhombifolia), Bigleaf maple (Acer macrophyllum), Pacific yew (Taxus brevifolia), and Oregon ash (Fraxinus latifolia) can be found in all channel locations. Higher on the channel bank, Douglas-fir and California bay laurel are common on terrace deposits. Unidentified mosses and sedges dominate the riparian understory. Elder Creek is a coarse grained channel with particle sizes ranging from 50 to 150 cm in steep riffles and 5 to 10 cm in point bar deposits. Bedrock is exposed on the floor of all major pools and comprises 10 to 90 percent of the banks in typical channel reaches. In Rosgen's c. Tops of isolated gravel deposits in the lee of midchannel boulders occurred at or below active stage; d. Presence of decomposed shale clasts (i.e. highly fractured but shape maintained) uncommon below active stage; e. Alluvial deposits at the downstream end of bedrock pools (the pools generally located at channel constrictions with no with no associated point bars) below active stage; f. A distinct bench in the cross section above the active channel crest (berm) created by a matrix of gravel packed in interstices of large cobbles and boulders, as in Figure 2. g. Mosses more abundant above active stage, but presence or absence of mosses on large cobbles and boulders not a reliable feature for active stage identification. The best field evidence of the active channel is often along the edge of coarse point or lateral bars. A sharp break in the cross section of a lateral bar occurs at the crest of the active channel (fig. 2B). A bench of coarse particles packed in a matrix of sand and small gravel

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

15

originates at the active channel crest and extends toward the bankfull stage. Osterkamp and Hupp (1984) call this bench the "channel shelf." Using the Trush, Connor, and Knight (1988) criteria, active and bankfull stage heights were identified at 18 cross sections. Stream discharges at the surveyed active and bankfull stage heights were calculated from discharge rating curves specific for each cross section. Average active discharge for all cross sections was 1.97 cubic meters per second (cms)(std. dev.= 0.29); average bankfull discharge was 14.7 cms (std. dev.= 1.82). An active channel capacity of 1.97 cms has an exceedance probability of 0.11 on the average daily flow duration curve, i.e., on the average, active channel discharge is equalled or exceeded 40 days of the year. A bankfull channel capacity of 14.7 cms has a recurrence interval of 1.55 years using the Log Pearson III distribution and 1.65 years with the lognormal distribution (U.S. Water Resources Council 1981), i.e., bankfull discharge is approximately equalled or exceeded one day in an average water year.

Methods
A detailed stream map was constructed for the lower 3.3 km of main channel in 1985 and 1986. A meter tape was set along the channel thalweg with compass bearings taken at each bend in the tape. Channel widths were measured with another tape placed perpendicular to the thalweg tape. Water surface slope was measured with an engineer's level during low summer flows. Features mapped included (1) variations in channel width at three stream discharges corresponding to bankfull, active, and mid-summer flows, (2) the channel thalweg (deepest location in a channel cross-section) at 2 m intervals, and (3) the dimension of all pools, riffles and alluvial bars. At map completion, measurement intensity averaged one set of channel width estimates for every 5 m of channel length. A map was drafted at a scale of 1:200. The locations of all riparian trees (those trees at the bankfull stage or lower) greater than 2 cm in diameter (d.b.h.) were plotted on the channel map during 1986 and 1987. Trees with their root bases positioned at, or closer to, the active stage height were considered part of the active channel riparian population. Trees with root bases located at, or closer to, bankfull stage were assigned to a bankfull population. All trees were observed during an active and a bankfull discharge. The stream channel was divided into homogeneous segments called planforms, utilizing the channel map. Riparian tree density (number of trees per meter of channel length) was calculated for each planform at active and bankfull stages. Criteria for determining the downstream extent of one planform and upstream edge of another planform included similarity in channel width, bank materials, curvature, and extent of depositional features. Cross-over of the channel thalweg from one bank to the opposite bank (i.e., at the inflection between two channel bends) provided an additional criterion for less confined planforms. Planform slope was calculated as the change in low flow water surface elevation over the thalweg length. Planform curvature was estimated as the change in thalweg direction. A straight line was drawn from the thalweg at the upstream edge of the planform to the thalweg at the apex of the bend. Another straight line was drawn from the thalweg's apex to the thalweg at the downstream edge of the planform. The change in direction, measured in degrees, from the upstream line to the downstream line was a simple measure of planform curvature. Using this procedure, the 3.3 km of channel was divided into 57 planforms, with slopes varying from 0.50 percent to 6.48 percent. Planform curvatures ranged from 1 to 110 degrees.

Figure 2 Photographs of Elder Creek riparian zone on a lateral bar (A) and close-up of the active channel (B). 16
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Results
Riparian populations at bankfull and active stages in the lower 3.3 km of the main channel of Elder Creek differed in species composition and relative abundance. A total of 581 trees were censused: 254 trees in the active population and 327 trees in the bankfull population. White alder and Oregon ash comprised 85.9 percent of the population at the active stage, while a more even mix of White alder, Oregon ash, Bigleaf maple, and Pacific yew comprised the bankfull population (fig. 3). Only a few individual alders and ash were noted above the riparian zone delineated by the bankfull stage. These individuals commonly were associated with springs or tributaries. The normally distributed size class distribution (size classes in d.b.h.) for the alder population and skewed size class distribution for ash (fig. 4) do not show abundant recruitment stocks of smaller individuals. Bigleaf maples were not limited to the riparian zone, but were common on steep slopes above bankfull stage. Pacific yew establishment ranged from the active channel berm up to terrace flats five meters above the bankfull stage. Size frequency distributions for maples and yews in Figure 4 only represent the riparian portion of their populations. Maple and yew recruitment also appeared low, particularly among yews with only two seedlings found along the channel. Riparian tree densities (number of trees per meter of planform length) at the active channel stage exhibited a threshold to changes in planform slope and curvature. We labeled this a threshold because tree densities spanned a range of values for a given planform slope or curvature but were bounded by an upper limit (called the threshold density). For example, tree density for the active population exhibited a threshold response to planform gradient (fig. 5). At a gradient of 1 percent, riparian densities ranged from 0 to 0.36 trees per meter, but usually did not exceed 0.25 trees per meter. The planform with 0.36 trees per meter was a braided channel reach. The only other braided reach also appears as an outlier in Figure 5, with a density of 0.14 trees per meter at a planform slope of 5.3 percent. Riparian density threshold at a slope of 1 percent was approximately 0.25 to 0.30 trees per meter; higher densities for non-braided channel reaches with slopes near 1 percent were not found. Threshold density decreased as planform slope increased, e.g. at a planform slope of 3.5 percent the threshold density was approximately 0.12 trees per meter. Threshold riparian density at the active stage also decreased with increasing planform curvature (i.e., greater degrees of bend). Threshold densities decreased from 0.25 trees per meter at a curvature of 5 degrees to a slightly lower density of 0.20 trees per meter at a greater

Figure 3 Relative species abundances of active and bankfull tree populations in the main channel of Elder Creek. curvature of 50 degrees (fig. 5). The change in threshold density was greater above 50 degrees, decreasing to 0.05 trees per meter for a 100 degree planform curvature. Changes in bankfull tree population density were not clearly related to planform slope and curvature. Threshold densities remained above 0.2 trees per meter over a wider range in planform slopes of 0.1 to 4.0 percent (fig. 6). A threshold density of approximately 0.20 trees per meter remained constant for planform curvatures ranging from 5 to 70 degrees (fig. 6). The use of graphical techniques does not allow separation of the relative importance of slope and curvature in determining threshold tree densities. We are not aware of a reliable statistical technique for defining a threshold response. The upper limit for riparian densities (approximately 0.40 trees per meter) probably results from the need for a minimum canopy area per mature tree rather than fluvial interaction. There must be a limit to the degree of "packing" trees in a confined space. In many planforms, riparian densities at the active or the bankfull channel was zero. Shear bedrock surfaces provided little opportunity for seedling establishment in several of the stream canyon reaches. The location at which a meandering channel changes direction (flowing from one meander into the next meander) is called a channel inflection point. At the inflection point, the thalweg is positioned near the center of the

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

17

Figure 4 Size frequency (d.b.h.) distributions for the four most common riparian species in the main channel of Elder Creek.

channel. All riparian trees found at the inflection (a 5 to 8 meter channel segment on both sides of the channel) were censused from the channel map. The average riparian density at bend inflections was 0.35 trees per meter. Average densities for the active and the bankfull stage at bend inflections were similar, 0.15 trees per meter and 0.20 trees per meter, respectively. Average tree density (combined active and bankfull populations) for the entire channel was 0.18 trees per meter, indicating that planform inflections provided relatively better habitat conditions for survival than other locations in the channel's planform geometry. Each tree was assigned to the inside or outside bank of its respective planform. In the active channel of Elder Creek, 121 trees occurred on the inside banks and 133 trees on the outside banks. In the bankfull channel, 185 trees were found on the inner and 142 trees on the outer banks. On the inner planform bank, the active riparian trees were less common than bankfull trees. On inner and outer banks of the active channel, riparian threshold density decreased with increasing planform curvature (fig. 7). Threshold riparian densities on the inner and outer banks of the bankfull channel did not exhibit as

clear a response to planform curvature (fig. 8).

Discussion
Riparian species must be physiologically adapted to periodic inundation and physically adapted to survive the forces of large storm flows. Our initial hypotheses centered on the expectation that greater exposure to storm flows reduced survival. Active channel riparian populations, established lower in the channel, should be more exposed to each storm's force than bankfull populations. Alluvial channels can adjust to major storm flows by overtopping the banks, thus increasing channel width to partially accommodate the flow. Flow depth in the main channel slightly increases above bankfull discharge. In contrast, the confined channel of Elder Creek, with no banks to overtop, accommodates storm discharge by increasing flow depth and/or velocity. If shear stress can be roughly approximated as the product of depth and slope, the inability to adjust channel width creates continually increasing shear on the stream bed and banks with larger storm flows.

18

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 5 Relationship between the density (trees per meter of planform) of the active channel tree population and planform slope and curvature. The solid line indicates the threshold riparian density. As a starting point, we hypothesized that steeper channels were more difficult places to survive, particularly for individual trees established on the active channel berm. The active population did exhibit a decrease in threshold density with increasing planform slope (fig. 5), while threshold densities for the bankfull population exhibited no clear trend (fig. 6). The two braided planforms, plotting above the threshold density curve in Figure 5, had bankfull widths twice the average of nonbraided planforms. Potential for width adjustment at high flows could reduce storm impact and favor riparian establishment.

Figure 6 Relationship between the density (trees per meter of planform) of the bankfull channel tree population and planform slope and curvature.

Lower planform curvatures were associated with higher riparian threshold densities, particularly for the active channel population. The most dense active populations of large alders occurred on short channel segments in planforms having slopes of 0.5 to 2.5 percent and low curvatures of 1 to 10 degrees; densities ranged from 0.25 to 0.40 trees per meter. Dense rows of alders occurred only on the active channel berm, often at planform inflections. The lowest active population densities were found on mobile point bars located on the inner bank of high curvature channel bends. On these bars, bedload movement occurred below bankfull discharge.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

19

Figure 7 Tree densities (number of trees per meter of planform) on the inner and outer banks of the active channel in relation to planform curvature. We only can infer process by examining patterns or trends. We hypothesized that the outer bank was a more harsh environment for riparian establishment because (1) vertical water velocity profiles would undergo significant change moving away from the bank at storm flows, thus creating high shear stress and (2) floating debris would be directed against the outer bank causing stem damage and uprooting. Channel-wide bankfull riparian abundance was slightly higher for the inner banks (185 trees) than the outer banks (142 trees). In many planforms, the outer banks had numerous crevices among the boulders and bedrock that provided sheltered sites for riparian establishment. Outer banks were composed of relatively nonerodible bedrock or boulders; bank undercutting was rare. Undercutting of the outer bank may not have the "opportunity" to be an important factor affecting population densities on Elder Creek. Other 20

Figure 8 Tree densities (number of trees per meter of planform) on the inner and outer banks of the bankfull channel in relation to planform curvature. stream types, with erodible banks, may have significant outer bank undercutting that greatly reduces survival to a large size for riparian species. Low densities of active trees (fig. 7), compared to higher bankfull densities on the inner bend (fig. 8), indicate another process (particularly on high curvature planforms). Most bedload movement through a channel bend at high flows did not occur along the thalweg, i.e., along the outer bend. Most movement occurred across the point bar. Seedlings on point bars (i.e., the inner channel bank) would encounter significant scour during bedload movement, as a flood exceeds active stage and covers the bars. The upper extreme of the bar surface marks bankfull or near-bankfull stage. Trees at this location would not be as exposed to bedload scour. The lack of fluvial "sand-papering" on exposed roots of
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

bankfull trees along the margin of point bars supports this hypothesis. Flood scarring of active channel trees appeared more frequent on the inner channel banks following a ten year storm in February, 1986. Riparian establishment along the active channel exists in close equilibrium with channel dynamics. In adjacent Rock Creek (fig. 1), channel width from the active channel crest to bankfull has expanded due to aggradation from intense logging beginning in the late 1940's. The aggradation induced straightening of the active and the bankfull channel. Even-aged alder stands colonized the outside portion of the original channel bends. On the inner banks of low curvature planforms, young alders often extend from bankfull stage down to the active stage. Following major storms and subsequent lowering of the channel bed (decreasing supply of bedload as Rock Creek watershed successfully revegetates), another row of younger alders becomes established along a new, narrower active channel. Lines of alders, each a different age, document episodes of channel adjustment. Alder pattern could be used to assess the rate of channel recovery for aggraded streambeds.

References
James, S.M. 1983. South Fork Eel River watershed erosion investigation. California Dept. Water Resources, Northern District, Redding, CA. 89 p. Osterkamp, W.R.; Hedman, C.R. 1977. Variation of width and discharge for natural high-gradient stream channels. Water Resources Research 13:256-258. Osterkamp, W.R.; Hupp, C.R. 1984. Geomorphic and vegetative characteristics along three northern Virginia streams. Bulletin of the Geological Society of America 95:1093-1101. Rantz, S.E. 1964. Surface water hydrology of coastal basins of Northern California. U.S. Geological Survey Water Supply Paper 1758, 77 p. Richards, K. 1982. Rivers: form and process in alluvial channels. London: Methuen. 358 p. Rosgen, D.L. 1985. A stream classification system. In: Riparian ecosystems and their management: reconciling conflicting uses. 1985 April 16-18; Tucson, AZ. Gen. Tech. Rept. RM-120. Fort Collins, CO: Rocky Mountain Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 91-95. Trush, W.J.; Connor, E.C.; Knight, A.W. 1988. Active and bankfull channel morphology of a bedrock stream. Department of Forestry and Natural Resources, University of California, Berkeley: 42 p. draft. U.S. Water Resources Council. 1981. Guidelines for discharge flood flow frequency. Washington, DC: Water Resources Council Bulletin 17B; 28 p. Williams, G.P, 1978. Bank-full discharge of rivers. Water Resources Research 14(6):1141-1154. Wolman, M.G.; Leopold, L.B. 1957. River floodplains: some observations on their formation. U.S. Geological Survey Professional Paper No. 282C:87-107.

Acknowledgements
The research was supported in part by the United States Department of the Interior, Geological Survey, through the State Water Resources Research Institute, Project No. CA-05-1984, the University of California Water Resources Center, Project UCAL-WRC-W-651, and funds provided by the Public Service Research and Dissemination Program (University of California). Contents of this publication do not necessarily reflect the views and policies of the U.S. Department of the Interior.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

21

THE MIDDLE SACRAMENTO RIVER: HUMAN IMPACTS ON PHYSICAL AND ECOLOGICAL PROCESSES ALONG A MEANDERING RIVER 1
Koll Buer, Dave Forwalter, Mike Kissel, and Bill Stohler 2

Abstract: Native plant and wildlife communities along Northern California's middle Sacramento River (Red Bluff to Colusa) originally adapted to a changing pattern of erosion and deposition across a broad meander belt. The erosion-deposition process was in balance, with the river alternately building and eroding terraces. Human-induced changes to the Sacramento River, including bank protection, gravel mining, pollution, riparian vegetation removal, flow regulation, and flood control, have resulted in a number of physical and ecological effects. This study focuses on changes in bank erosion, bank composition, river length, depth, width, and sinuosity, and floodplain deposition (ongoing study, Sacramento River Bank Erosion Investigation, Department of Water Resources, Red Bluff, California.) The Department of Water Resources, Northern District, is monitoring these changes using old survey maps, aerial photographs, and field surveys. Completed studies indicate that bank protection has significantly reduced a source of salmon spawning gravel from freshly eroded banks and will over time decrease the number of preferred spawning areas such as point bar riffles, chute cutoffs, multiple channel areas, and areas near islands. Bank protection also increases the tendency of the confined river to deepen and narrow, further reducing spawning habitat. Because of flood protection from dams and extensive bank protection along eroding banks, most of the rich high terrace soils and all but a few percent of the original riparian forest has been converted to agriculture and other uses. In addition, only 45 percent of the original streambank vegetation remains. Wildlife populations have declined markedly due to loss of riparian habitat and suppression of the natural successional processes that maintain the density and diversity of habitat within the riverine environment. Some species that are adapted to the dynamics of the erosional-depositional cycle are threatened with extinction or extirpation as key habitat elements are lost from the newly stabilized river system. Flood control has interrupted the natural equilibrium between erosion and deposition, resulting in reduction in bank erosion rates and in overbank sediment deposition. Solutions to these problems will require a comprehensive river management program that incorporates the natural processes of meandering and bank erosion.

The Sacramento River is the largest and most important river system in California. The basin represents about 17 percent of California's land area, yet yields 35 percent of the water supply. It drains the north half of the Great Central Valley of California. The river is the State's most important salmon resource. Biologically, the riparian corridor between Red Bluff and Colusa is also one of the richest and most diverse that remains in California. The Department of Water Resources (DWR), Northern District, has an ongoing study of geomorphic changes in this vital reach of the river (fig. 1). The Department began its bank erosion studies in 1977. For two years, four bank erosion sites were monitored. A report was published in 1979 outlining the results (DWR 1979b). Several of these sites were monitored until 1983. A new monitoring program began in 1986. Nine new bank sites were surveyed in the 58-mile study reach. An additional six sites were surveyed in 1988. The scope was also expanded to include overbank deposition and bank composition. Ten floodplain cross-sections were surveyed to monitor sediment deposition and longterm changes. The study also considers long-range geomorphic changes caused by such human activities as dams, bank protection, and gravel mining. Geomorphic changes include channel narrowing and deepening; changes in riparian vegetation, channel length, width, sinuosity, and bank erosion, and sediment transport rates. Historical Aspects Before settlement of the Sacramento Valley, the Sacramento River was free flowing. Late summer flows were low, averaging 3,000 cubic feet per second (cfs), and in dry years dropping as low as 1,000 cfs. The river, however, would fluctuate widely in response to winter rains and spring snowmelt. Periodically, it would overflow its banks and flood large areas of the valley floor. These areas were covered by dense forests of riparian vegetation adapted to the periodic flooding.

1 Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

2 Senior Engineering Geologist, Graduate Student Assistant, and Student Assistants, respectively, Department of Water Resources, Red

Bluff, California.

22

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 Sacramento River from Keswick Dam to Colusa

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

23

Below Red Bluff, bank erosion and lateral migration across the floodplain were natural processes. Large floods uprooted streamside vegetation, caused banks to recede and the river to meander. Sediment derived from tributaries and from bank erosion deposited in overbank areas where vegetation reduced water velocities. Over a period of years, erosion and deposition were roughly in balance, so that the valley floor neither aggraded nor degraded. The riparian forests played a doubly important role here, first by reducing bank erosion and, second, by inducing deposition on the floodplain. A large number of chinook salmon (Oncorhynchus tshawytscha) migrated up the Sacramento River each year to spawn. Although there were probably four runs then, as there are today, the two largest runs were thought to have occurred in the fall and spring. The other two runs, winter and late fall, are not as well documented historically, especially their numbers. Most of the spring- and winter-run salmon, as well as part of the fall and late-fall salmon, were thought to have spawned upstream from the present location of Shasta Dam. However, large numbers of spring- and fall-run salmon also spawned in many Sacramento tributaries. Since about 1850, the study reach has undergone a number of hydrologic, geomorphic, and environmental changes, most of which have been detrimental to locally adapted species. These changes are caused by dams and diversions, bank protection, urbanization, stream gravel removal, hydraulic mining, agriculture, and logging. Many of these changes have had long-reaching effects, including alteration of river characteristics, such as depth, width, gradient, sinuosity, and bank erosion. This in turn has reduced riparian vegetation, water quality, hydrologic diversity, and fish and wildlife resources. Urbanization, primarily in Redding, Anderson, Cottonwood, and Red Bluff, has caused additional problems in the study reach. Gravel extraction for highways, housing, and other projects averages more than 1.3 million cubic yards per year in Shasta County and 0.5 million in Tehama County, mostly from tributary streams. This, in conjunction with Shasta, Keswick, Whiskeytown, and other dams that prevent gravel recruitment from upstream reaches, has eliminated the spawning gravel available in downstream reaches. Along with the rapid expansion of the mining industry, California agriculture also grew. First to be converted to agriculture were the fertile rimlands. Rimlands are next to the river, higher than the surrounding tule lands, and are less often flooded. Flood control had its inception in the low levees constructed on the rimlands by farmers protecting their crops. Next to be developed were the tule, or swamp and overflow, lands. Through a series of legislative acts 24

passed between 1855 and 1868, the State sold these lands to farmers who were obliged to reclaim them individually or through the formation of reclamation districts. Within a period of 3 years following the last act, practically all of such lands had passed into private ownership (Jones 1967). To date, as a consequence of just these two kinds of agricultural development, about 98 percent of the original riparian forest has been removed (DWR 1988). Dams and unscreened or poorly screened diversions have severely depleted the river's fishery. Early dams and diversions built by miners and farmers obstructed miles of habitat without allowance for fish passage or mitigation measures. By the 1920s, at least 80 percent of the Central Valley spawning grounds had been cut off by obstructions, according to the U. S. Bureau of Reclamation (USBR 1972). Dams affect riparian areas mostly by the reduced incidence of flooding, bank erosion, and silt deposition required for the regeneration of riparian habitat. Flood control also encourages the development of riparian lands along the river. More recently, major water development projects, such as Shasta and Keswick Dams and the Trinity River Diversion, have affected Sacramento River hydraulics. Shasta Dam stores 4.5 million acre-feet and, to a large extent, regulates flows from the Pit, McCloud, and upper Sacramento Rivers. Keswick Dam, 9 miles downstream from Shasta provides water regulation, stops salmon migration and acts as a fish-trapping facility. The effect of Shasta Dam on the natural flow (DWR 1984) has been to: 1. Decrease the minimum discharge and increase the number of very low discharges. This occurs when the powerhouse is closed for repairs. 2. Increase the number of moderate discharges, particularly during the summer and fall irrigation season. 3. Reduce the number and the volume of high and very high flows. Since December 1963, water has been diverted from the Trinity River Basin through the Clear Creek Tunnel and Judge Francis Carr Powerhouse to Whiskeytown Lake. The Spring Creek Tunnel then diverts Trinity water and most of Clear Creek water through another power plant into Keswick Lake. An average of about 1 million acre-feet of Trinity River water is now diverted into the Sacramento River Basin each year. The effect of the Trinity River diversion on postShasta flows has been to increase average Sacramento River discharge by about 1,000 to 1,500 cubic feet per second throughout most of the year.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

River Geomorphology and Ecology

In alluvial river systems, banks erode and sediments are deposited. Floodplains, islands, and side channels will undergo modification with time. Bank erosion generally occurs on the outside of meander bends. Here, banks are susceptible to erosion because high-flow velocities impinge directly onto banks and turbulent motion along the channel thalweg undercuts the banks. Eroding banks may be either highterrace or low terrace. High terrace banks normally have a deep soil profile containing mostly loamy sand and silt. Below the soil is a deposit of sand and gravel which usually originates as the point bars on the wide edges of river beds. A low-terrace bank consists mostly of a sand and gravel which was deposited as point bars with a thin soil profile on top. The fish, wildlife, and riparian vegetation are adjusted to the cycle of erosion, deposition, and changing channel pattern in which the river swings slowly back and forth across a broad meander belt. The health and productivity of the system at any one point is dependent on the periodic rejuvenation associated with these changes. Salmon prefer to spawn in fresh, uncompacted gravels that have recently moved. These spawning beds tend to occur in wide areas with multiple channels or chute cutoffs because of increased spawning flow velocity, reduced flood-flow velocity, shallower depths, and greater hydraulic diversity there. Gravel in the subsoil horizons of an eroding bank provides fresh gravel to spawning beds. Much of the sand and silt from the bank is redeposited in the riparian forests downstream. Abandoned channel oxbows and sloughs are ideal habitat for warmwater fish, such as largemouth bass and catfish. Bank erosion is important for the recruitment of spawning gravel. It allows for the re-entrainment of gravel deposited on point bars. Between Red Bluff and Colusa, bank erosion is estimated to contribute about 85 percent of the total available spawning gravel (DWR 1984; 1985). Bank erosion is also the driving force for riparian plant succession. On the outside of bends, high-terrace banks with a mature forest typically consisting of valley oak, box elder, and black walnut are eroded. On the opposite side is a point bar consisting of sand and gravel. Willows and cottonwoods become established here. The rapid invasion of riparian vegetation slows floodflow velocities and allows sand and silt to deposit. With time, riparian stages with a succession of different plant species occurs as the point bar becomes higher and farther away from the river. Various birds and other wildlife use different riparian stages for feeding, nesting, and reproduction. The climax valley oak forests are relatively sterile compared

Using such channel characteristics as gradient, geometry, underlying rock types, and gravel distribution, it is possible to divide the Sacramento River between Redding and Colusa into seven distinct reaches. These reaches were described in detail by the Department of Water Resources (1984; 1985) and are only briefly described here. Typically, the river between Redding and Red Bluff (reaches 1 to 5) is a bedrock reach. The river is entrenched, in places, with some vertical banks 100 feet or more high. Below Red Bluff, the Sacramento River is mostly an alluvial stream. Reach 6 is between Red Bluff and Chico Landing. It is sinuous and anabranching. Reach 6 has been divided into eight subreaches (6A to 6H) based on bank erosion, sinuosity, and meander belt width. Reaches 6A, 6C, 6E, and 6G are short, narrow, straight reaches with low sinuosity, low gradient, and only minor bank erosion. Between the short, stable reaches are longer, more sinuous, unstable reaches 6B, 6D, 6F, and 611. Reach 7 is between Chico Landing and Colusa. Here the gradient is less; the river tends to be more sinuous with fewer islands. The most distinctive feature of this reach is the gradual downstream development of natural levees. This reach has been divided into five subreaches (7A to 7E), based on the criteria used for reach 6. Bank Erosion A river erodes both its banks and bed. Bed erosion leads to degradation and grading of the stream profile. In a bedrock stream, this process is generally slow, except during periods of geologically rapid rejuvenation and uplift. Bed erosion also occurs in alluvial streams, but the erosion is generally balanced by deposition over a period of years. Bank erosion is generally of much more interest and concern to people. Bank erosion is dependent on channel shape, bed and bank material, and river hydraulic characteristics. Because of the generally stable banks of the Sacramento River between Keswick and Red Bluff, bank erosion is insignificant in most places. Between Red Bluff and Colusa, significant bank erosion occurs. Downstream of Colusa, flows and associated velocities are greatly reduced by overflow occurring upstream during flooding (both overbank flow and flow at the Moulton and Colusa overflow weirs). In addition, the flatter slopes of the channel bed downstream minimize the erosion potential.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

25

to the younger riparian stages. Therefore, bank erosion and riparian rejuvenation are necessary to maintain a healthy and productive ecosystem. Sediment deposition is inextricably linked to bank erosion. Without deposition, the channel would simply widen until it was so large that erosion would terminate. However, the coarser material eroded from the bank is deposited on point bars downstream. The point bars constrict the bend and enable erosion to continue. DWR (1979) observed bank erosion over a 2-1/2-year period at six sites in the Red Bluff-to-Colusa Reach. This investigation has been expanded and is presently continuing. Bank erosion was divided into summer (lowflow) erosion and winter (high-flow) erosion. Only two of the six sites showed any erosion during the summer. Average bank recession between April and October 1977 was 11.4 and 2.2 feet, respectively. In contrast, high flows were far more conducive to erosion. A major storm occurred in January 1978. Erosion was greatest during the period that included this storm, with erosion ranging from 30 to 50 feet of bank recession. During the storm itself, Woodson Bridge State Recreation Area below Thomes Creek lost over 40 feet in a single 24-hour period. Beginning in 1986, nine new erosion sites were surveyed. Six more were added in 1988, for a total of fifteen sites. Figure 2 shows one of these. Each site is surveyed three to four times yearly. Successive bank lines are plotted and eroded bank area calculated. Figure 3 shows the Sacramento River discharge and bank erosion between 1986 and 1988 at ten erosion sites. Note that no floodflows occurred during the 2 years and that the highest peak flow was about 65,000 cfs. At Ord Ferry, the flood stage, or bankfull discharge, is about 110,000 cfs and floodflow in March 1983 was about 160,000 cfs. Six of the nine sites had essentially no erosion during the low-flow period. The Princeton site has eroded several thousand feet since 1978 and is the most erodible site on the river. It eroded an average of 4 feet during the low-flow period and an additional 30 feet during the two higher flow periods. Another goal of the bank erosion study was to determine total bank erosion, gravel and silt produced from

bank erosion, and bank recession rates. A total of 67 bank erosion sites were identified and evaluated by comparing 1976 and 1987 aerial photographs. Eroded areas were measured using a planimeter. Thirty of the sites were field surveyed. Longitudinal profiles were measured showing bank height, thickness of gravel, and thickness of silt. Gravel samples were sieved and analyzed. The additional 37 sites were identified as showing significant erosion, but were not sampled or measured in detail. Data from the surveyed sites were averaged by geomorphic reach and applied to the unsurveyed sites. Aerial photographs from 1976 and 1987 were compared and bank erosion areas measured using a planimeter. Table 1 is a summary of bank erosion data for the reach. There are a number of interesting observations that can be made from this table. The average bank height from the bottom of thalweg to the top of the bank is about 25 feet, with 16 feet of gravel and 9 feet of silt. The average bank recession of an actively eroding bank is about 20 feet per year, ranging from a few feet to nearly 80 feet. The average length of an eroding bank is about 3,000 feet, for a total of about 197,000 linear feet. This represents 34 percent of the bank in the reach. Another 18 percent was eroding but has been protected with rock riprap. Between 1976 and 1987, about 1,230 acres have been eroded. This has produced about 29 million cubic yards of sand and gravel and 20 million cubic yards of silt. Most of the gravel is redeposited on point bars immediately below the eroding bank. Some of the silts are deposited on the floodplain and some are washed downstream. DWR (1979a,b), U. S. Corps of Engineers (USCE 1981), and the U. S. Geological Survey (USGS 1977) have compared pre- and post-Shasta erosion rates. All three investigations concluded that there has been a significant reduction of about 25 percent in bank erosion between the period 1896-1946 and 1946-1980. The differences in rates can be attributed in part to a reduction in the frequency and magnitude of peak flows resulting from regulation by Shasta Dam. Secondly, since bank erosion increases exponentially with discharge, any reduction in the occurrence of high flows will reduce the amount of bank erosion (DWR 1984; 1985).

26

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 2 Golden State Island Surveyed Erosion Site December 9, 1986 through June 7, 1988

27

28
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 3 Sacramento River Discharge and Bank Erosion for Three Hydrologic Periods - December 9, 1986 through June 7, 1988

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 1. Sacramento River Erosion Site Parameters


29

Between Red Bluff and Chico Landing, 94,000 feet of riverbank has been riprapped and an additional 81,000 feet of riprap have been proposed. If this is developed, the total riprap in this reach would comprise 34 percent of the riverbank. In the Butte Basin reach (River Mile 176 to 194), 15 percent of the banks are protected and an additional 10 percent are planned (USCE 1988). Bank protection, when effective, stops bank erosion and lateral migration. It prevents loss of valuable agricultural lands, transportation facilities, and structures. Bank protection, particularly if it is along all the eroding banks of the river, will cause some long-range geomorphic changes. First, it will have a stabilizing effect on length and sinuosity. Second, it will prevent the re-entrainment through bank erosion of gravel deposited on point bars. This will have some long-range effects on the amount of available spawning gravel. Third, over a period of time, it will tend to narrow the channel, increase the depth of flow, and reduce the hydrologic diversity. Sloughs, tributary channels, and oxbow lakes will fill with sediment and no new ones will be created. Changes in Length and Sinuosity Analyses of channel length and sinuosity were done on eleven sets of maps and photographs dated between 1896 and 1987. These were tabulated and published in DWR (1985). No trends were apparent, however, in that some reaches are increasing in length and sinuosity and others are decreasing. Changes in Depth and Width The Department of Water Resources (1984) theorized that bank protection would cause the channel to narrow and deepen. When a channel is stabilized, it will no longer erode its banks but will erode its bed. Deposition will continue on the inside of the bend, causing the channel to narrow and deepen. Such a channel will have less hydraulic diversity and salmon spawning area. As part of the bank erosion studies, sonar surveys were used to measure depths along eroding and riprapped banks. The surveys were done during the summer of 1987. River widths were measured from June 1987 aerial photographs. Figure 4 shows the thalweg depths plotted with river mile. There were 37 riprapped banks and 28 eroding banks measured and plotted. The lines are 4-point running averages showing that the mean thalweg along riprapped banks average 6 feet deeper than comparable eroding banks. A similar analysis was completed for river widths. Figure 5 shows that the average width at riprapped sites is about 90 feet less than at eroding sites. The difference in widths appears to remain fairly constant from Ord 30

Ferry (River Mile 184.3) to River Mile 223; then the difference decreases until from River Mile 235 to Red Bluff, it is essentially nonexistent. This is a consequence of the more stable banks near Red Bluff. The effects of Shasta Dam on river geomorphology are complex. In principle, Shasta Dam would tend to reduce width because of less frequent floodflows. However, this effect may be offset by factors that tend to increase channel width, such as riparian loss. Further study is necessary to adequately assess this complex issue.

Conclusions
Some of the conclusions of the study are that: Human activities causing adverse impacts include hydraulic gold mining, gravel mining, bank protection, flood control, and removal of riparian vegetation by agriculture. Changes in the hydrologic regime, primarily due to Shasta Dam, have reduced bank erosion rates and overbank deposition. Eroding banks consist predominantly of gravel, and the gravel is vital in replenishing salmon spawning areas. Floodplain deposition replaces silt lost by bank erosion. Our ongoing study will help to determine if the system is in balance. Bank erosion is necessary to maintain riparian succession, anadromous fish and wildlife habitat in the reach. Most of the bank erosion occurs during the winter. Summer erosion is significant in a few places along the river. Eroding sites differ considerably in bank erosion rates. The average rate is about 20 feet per year for the last 11 years. Between Red Bluff and Ords Ferry, this is about 110 acres per year or about 2 acres per year per mile. The high rate is most likely due to two large floods that occurred during this period. Bank protection will have several negative, longrange effects. Since over 85 percent of the gravel provided to the river comes from bank erosion, it will prevent the recruitment of gravel to salmonspawning areas. It will cause the river to narrow and deepen. A survey of riprapped and eroded banks shows that riprapped banks are an average of 6 feet deeper at the thalweg. A similar survey of river width shows that the river at riprapped banks is an average of 90 feet narrower. Finally, it will affect the natural rejuvenation of riparian vegetation, reducing or eliminating the habitat for many species of birds and animals.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 4 Thalweg Depths Next to Eroding and Riprapped Banks Sacramento River, Summer 1987

Recommendations
Recommendations of the study include development of a comprehensive management plan that will: Protect, enhance, and maintain the natural fish, wildlife, and riparian habitat along the Sacramento River. This includes regulation to protect erosiondependent features such as salmon spawning areas, bank swallow nesting sites, oxbow lakes and offstream wet areas, riparian vegetation, and the attendant wildlife habitats. Continue monitoring programs of the Sacramento River geomorphology, including bank erosion, meandering, river depth and width, spawning gravel supply, and floodplain deposition.

References
California Department of Fish and Game. 1965. "California Fish and Wildlife Plan". Volume III, Part B. Sacramento, Calif. CDFG. 370 p. California Department of Water Resources, Northern District. 1979a. "Land use change in the Sacramento River riparian zone, Redding to Colusa, An update - 1972 to 1977". Red Bluff, Calif. 34 p. California Department of Water Resources, Northern District. 1979b. "Observations of Sacramento River bank erosion, 1977-1979". Red Bluff, Calif. 58 p. California Department of Water Resources, Northern District. 1980. "Upper Sacramento River spawning gravel study". Red Bluff, Calif. 200 p. California Department of Water Resources, Northern District. 1983. "Land use changes in the Sacramento River riparian zone, Redding to Colusa, A second update - 1977 to 1982". Red Bluff, Calif. 28 p. California Department of Water Resources, Northern District. 1984. "Middle Sacramento River spawning gravel study." With River Atlas. Red Bluff, Calif. 194 p. California Department of Water Resources, Northern District. 1985. "Sacramento River spawning gravel studies Executive Summary." Red Bluff, Calif. 38 p.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

31

California Department of Water Resources, Northern District. 1987. "Land use changes in the Sacramento River riparian zone, Redding to Colusa, A third update - 1982 to 1987". Red Bluff, Calif. 20 p. Jones, Gerald II. 1967. Alteration of the regimen of the Sacramento River attributable to engineering activities during the past 116 years. Sacramento Section, U. S. Corps of Engineers. United States Bureau of Reclamation (USBR). 1972. "Salmon in California". United States Corps of Engineers. 1978. "Reconnaissance report Sacramento River and tributaries, bank protection and erosion control investigation, California". Sacramento, Calif. 94 p. United States Corps of Engineers. 1981. "Sacramento

River and tributaries bank protection and erosion control investigation, California". 252 p. United States Corps of Engineers. 1988. "Butte Basin Reach - Sacramento river bank protection project". Sacramento, Calif. 211 p. United States Department of the Interior, Bureau of Fisheries (now U. S. Fish and Wildlife Service). 1940. "An investigation of fish salvage problems in relation to Shasta Dam". Special Scientific Report No. 10. Washington, D. D.C. United States Geological Survey. 1977. "Lateral migration of the Middle Sacramento River, California". U. S. Geological Survey Water Resources Investigations 77-43. 51 p.

Figure 5-- River Widths Next to Eroding and Riprapped Banks Sacramento River, Summer 1987

32

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

INFLUENCE OF CHANNEL GEOMORPHOLOGY ON RETENTION OF DISSOLVED AND PARTICULATE MATTER IN A CASCADE MOUNTAIN STREAM1
Gary A. Lamberti, Stan V. Gregory, Linda R. Ashkenas, Randall C. Wildman, and Alan D. Steinman 2 Abstract: Retention of particulate and dissolved nutrients in streams is a major determinant of food availability to stream biota. Retention of particulate matter (leaves) and dissolved nutrients (nitrogen) was studied experimentally during summer 1987 in four 300-500 m reaches of Lookout Creek, a fifth-order stream in the Cascade Mountains of Oregon. Constrained (narrow valley floor) and unconstrained (broad valley floor) reaches were selected within old-growth and second-growth riparian zones. Gingko leaves and ammonium were released into the channel and retention rates were measured. Retention of leaves and nutrients was 2-4 times higher in unconstrained reaches than in constrained reaches, in both old-growth and second-growth riparian zones. Retention was enhanced by increased geomorphic complexity of channels, diversity of riparian vegetation, presence of woody debris, and heterogeneity in stream hydraulics, sediments, and lateral habitats. Unconstrained reaches express these qualities and thus are critical areas for retention of particulate and dissolved nutrients in stream ecosystems. tors can influence retention rates by modifying either the physical structure or the biological organization of stream ecosystems. In this paper, we focus on stream channel geomorphology as an important parameter regulating the retention of dissolved and particulate matter in streams. However, we also report on how the retention process is influenced by the interaction between channel geomorphology and other factors such as landform, riparian vegetation, and hydraulics. We used an experimental approach involving releases of leaves and nutrients into the channel to determine rates of retention in different geomorphic settings.

Study Area
Releases of leaves and nutrients were conducted in Lookout Creek, a fifth-order stream located within the H.J. Andrews Experimental Forest in the Cascade Mountains of Oregon, USA. We studied four discontinuous reaches in Lookout Creek (designated C, D, E, and G in this study), ranging from 336 - 547 m in length. Elevations of the study reaches ranged from 400-600 m, channel gradient was 2-3 percent, and summer baseflow discharge was 0.28-0.42 m3/sec. Stream substrates were dominated by cobble and small boulders. Accumulations of large woody debris in the channel were significant features of some reaches. The four reaches were classified on the basis of valley floor landform and riparian vegetation (Table 1). The valley floor is considered to be that portion of the valley that has been influenced by fluvial processes. Two categories of valley floors represented in this study were: constrained reaches, in which the valley floor was less than 3 times the width of the active stream channel, and unconstrained reaches, in which the valley floor was greater than 4 times the active stream channel width (Fig. 1). Based on these criteria, reaches C and E were constrained and reaches D and G were unconstrained. Riparian vegetation was classified as old-growth conifer forest or second-growth mixed deciduous and conifer forest. Reaches C and D (one constrained and the other

Stream ecosystems face the fundamental challenge of retaining the dissolved and particulate nutrients delivered to them by their watersheds. Because the biota of many streams is dependent on input of nutrients from riparian zones (Vannote and others 1981), retention of these materials is essential to lotic food webs. Streams differ markedly from other ecosystems in that the unidirectional flow of water tends to transport nutrients to downstream reaches. As a consequence, mechanisms for retention of dissolved and particulate matter must be present within the stream reach before nutrients can be utilized effectively by the biota (Elwood and others 1983; Speaker and others 1984; Young and others 1978). The process of retention thus provides a critical linkage between nutrient input and biotic utilization in lotic ecosystems. The premise of our research is that retention of dissolved and particulate matter is a function of valley floor landform, riparian vegetation, channel geomorphology, and stream hydraulics and substrate. Each of these fac-

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

Assistant Professor of Fisheries, Associate Professor of Fisheries, Department of Fisheries and Wildlife, Oregon State University, Corvallis, Oregon; and Postdoctoral Research Associate, Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tenn.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

33

Table 1 General characteristics of the four study reaches in Lookout Creek, Oregon. Reach Parameter C D E G Valley Floor Constrained Unconstrained Constrained Unconstrained Riparian Vegetation Old-Growth Old-Growth 2nd-Growth 2nd-Growth Multiple Channels Absent Present Absent Present Sediments Shallow Deep Moderate Deep Canopy Closed Partly Open Open Partly Open Lateral Stream Habitat Limited Abundant Limited Abundant Woody Debris Sparse Abundant None Abundant Dominant Channel Unit Pool/Cascade Riffle/Rapid Rapid Pool/Rapid

Figure 1 Schematic representation of constrained reach C and unconstrained reach D in Lookout Creek.

unconstrained) had old-growth riparian vegetation dominated by mature Douglas fir (Pseudotsuga menziesii), western hemlock (Tsuga heterophylla), and western red cedar (Thuja plicata). The other two reaches E and G were in second-growth riparian zones dominated by young stands of red alder (Alnus rubra), big-leaf maple (Ater macrophyllum), and some conifers. Mapping of valley floor landforms and channel geomorphology was conducted prior to this study (G. Grant 34

and F. Swanson, unpub. data). The unconstrained reaches had two channels that spanned most of the reach length, whereas the constrained reaches consisted of a single channel. Specific reaches were divided into channel units, which were identifiable geomorphic components of the channel at least as long as the channel was wide. Channel units in this montane stream included cascades, rapids, riffles, and pools.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Materials And Methods

Retention can be thought of as the difference between the quantity of particles or dissolved nutrients in transport at a given point in the stream and the quantity still in transport at some point downstream, given no significant new inputs over that distance. This difference is best expressed in some standardized form because quantities of released particles or nutrients, length of stream examined, or duration of measurement will vary from experiment to experiment. Previous releases of leaves indicate that retention is represented well by a negative exponential model (Speaker and others 1984): Td =Tie-kd where Ti is the total amount of released leaves or nutrients, Td is the amount of leaves or nutrient in transport at some distance d downstream of the release point, and k is the instantaneous rate of removal from transport (instantaneous retention rate or retention coefficient). The average travel distance of a leaf or nutrient atom is the inverse of the retention coefficient, or 1/k (Newbold and others 1981).

Leaf Retention

Figure 2 Empirical models of leaf and nutrient retention in streams. A negative exponential model provided the best fit for observed retention of leaves and ammonium.

Ginkgo leaves (Ginkgo biloba) were released into the channel to measure particulate matter retention. Ginkgo leaves are about the same size as leaves of many riparian tree species, but gingko trees do not occur naturally in North American riparian zones. The leaves are bright yellow and thus easily distinguished from those of other riparian species. Ginkgo leaves were collected just after abscission, dried, and counted into 5000-leaf batches. Leaves were soaked in water for 12 hours prior to release to ensure neutral buoyancy in the stream. One batch of 5000 leaves was released at the top of each study reach. At 100 m downstream, a net was placed across the stream channel to capture any leaves that were not retained within that 100-m segment. All leaves caught by the net for 2 hr after the release were retrieved and counted. Leaf retention downstream of the release point conformed to a negative exponential model in most cases (Fig. 2).

Nutrient Retention Because ambient nutrient levels throughout Lookout Creek typically are about 15 ug/L total inorganic nitrogen (TIN = NO3-N + NO2-N + NH4-N) and 15 ug/L soluble reactive phosphorus (SRP = PO4-P) for a N:P ratio of 1:1, production is usually limited by the availability of nitrogen. Ammonium nitrogen (as NH4C1) was released into Lookout Creek to measure dissolved nutrient retention. NH4Cl, the labile nutrient, was mixed in water with rhodamine-WT, a conservative dye tracer, to form a concentrated batch of nutrient and dye. This mixture was released at a constant rate into a turbulent area of the stream using a steady-head dripper system. Dye and nutrient usually mixed completely with stream water within 50 m of the release point. Sufficient ammonium was released to raise equilibrium TIN levels to between 50 and 70 ug/L (i.e., 3-5X above ambient) at

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

35

the upstream end of the reach. When equilibrium was reached, two water samples were taken at the bottom of each channel unit within the reach. Dye concentration in one sample was measured on a Turner fluorometer. Ammonium concentration was determined in the other sample with a Technicon II autoanalyzer using the phenate method. The conservative dye tracer provided a measurement of downstream dilution due to incoming tributaries, springs, seeps, and groundwater (Fig. 2). All nutrient measurements were adjusted for dilution using the correction NH4-Nd . (Dyei/Dyed), where i is the concentration at the top of the reach and d is the concentration at some distance d downstream where the corresponding nutrient sample was taken. Dye releases also allowed measurement of discharge and hydraulic retention. All releases of leaves and nutrients were conducted at summer baseflow (0.3-0.4 m3/sec) during August and September of 1987.

Table 2. - Geomorphic and hydraulic characteristics of the four study reaches in Lookout Creek, Oregon. Reach Parameter C G D E 26.0 Valley Floor Index1 1.3 6.9 2.8 Thalweg Length (m) 410 547 349 336 415 1198 391 663 Total Length3 (m) Mean Width of Active 18.7 28.0 26.3 230.0 Channel (m) 3948 7080 3005 3597 Active Channel Area (m2) No. Channel Units 17 52 13 34 0.31 0.40 0.28 0.31 Discharge (m3/sec) Reach Saturation 1:45 3:30 0:55 1:50 Time (h:m) 1Walley floor width (m) / Active channel width (m) 2Estimated 3Cumulative length of all channels Table 3 - Leaf and nutrient (NH4) retention in the four study reaches of Lookout Creek, Oregon. Length of experimental reach was 100 m for all leaf releases, and thalweg length for all nutrient releases. Reach C D E Parameter G Leaf retention Coefficient k 0.008 0.031 0.008 0.045 Average Leaf Travel 132 33 125 22 Distance (m) NH4 Retention Coefficient k .0014 .0092 .0059 .0044 Average Ammonium Travel 732 109 171 229 Distance (m) 6.0 15.3 17.8 25.8 Absolute Retention (ug NH4-NL-1.100m-1)

Results
Geomorphology and Hydraulics The valley floor index, which is the ratio of the valley floor width to the active stream channel width, ranged from 1.3-2.8 in the constrained reaches C and E to >6.0 in the unconstrained reaches D and G (Table 2). Unconstrained reaches had 2-3 times the number of channel units as constrained reaches. This is due, in part, to the two parallel channels present in unconstrained reaches, which contributed units, wetted surface area, and total length to those reaches. For example, thalweg length was similar in the old-growth pair C and D and the second-growth pair E and G, but total length was two to three times higher in the unconstrained reach of each pair. Further, active channel area in unconstrained reach D was almost double that of constrained reach C. This a real difference was not as striking in the secondgrowth reaches E and G because constrained reach E was located on an alluvial fan, which resulted in a somewhat broader channel. However, downcutting through the fan still applied local constraint to reach E. Discharge was 10-30 percent higher in unconstrained reaches than in constrained reaches, probably because of flow contributions from large stores of subsurface water in the deeper sediments of those reaches (Table 2). The time required for dye saturation of unconstrained reaches was about twice as great as for constrained reaches, indicating greater hydraulic residence time in the unconstrained reaches. In other words, unconstrained reaches retained water for a much longer time than did constrained reaches.

Leaf Retention Leaf retention coefficients ranged from 0.008 in both constrained reaches to 0.031-0.045 in the unconstrained reaches (Table 3). The average travel distance of an individual leaf was 125-132 m in the constrained reaches compared to only 22-33 m in the unconstrained reaches. In constrained reaches, slightly over 50% of released leaves were retained in the 100-m reach whereas >95% of leaves were retained in the unconstrained reaches. The riparian setting (old-growth or second-growth) did not appear to influence rates of leaf retention, but rather valley constraint was the predominant factor. Within specific reaches, high-velocity channel units with high roughness, such as rapids and cascades, retained more leaves than lower-velocity units with limited roughness, such as pools. Units with significant amounts of bedrock trapped few leaves. We observed that most retained leaves were trapped in interstitial areas between rocks, especially cobble and boulders. Substrates along stream margins trapped leaves more efficiently than similar substrates in mid-channel areas. Sticks and branches, where present, trapped leaves with high efficiency. Accumulations of woody debris, which occurred along stream margins in Lookout Creek, also trapped large quantities of leaves. Thus, in general, stream margins were more retentive than mid-channel

36

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

areas, especially when these lateral habitats had roughness elements such as protruding rocks or woody debris. Nutrient Retention Ammonium retention coefficients ranged from 0.0014 in the old-growth constrained reach to 0.0092 in the old-growth unconstrained reach (Table 3). Retention coefficients were intermediate and similar in the secondgrowth reaches. Nutrient retention was very low in the constrained old-growth reach, where the average nitrogen atom traveled over 700 m before being retained; in absolute quantities, only 6 ug/L of NH4-N (or 10.9 percent of that released) was retained over 100 m. In contrast, the average nitrogen atom traveled only 109 m in the unconstrained old-growth reach, and absolute uptake was over twice that of the constrained reach. Over 25 percent of the released ammonium was retained in 100 m. In the second-growth reaches, nutrient retention was slightly higher in the unconstrained reach, where about 26 ug/L of released nitrogen was retained in 100 m. However, nutrient retention was quite high in the constrained reach as well (18 ug NH4-NL-1.100 m), in particular when compared to the constrained reach in the old-growth. Over one-third of the released nutrient was retained in 100 m of the unconstrained reach, and about 25 percent was retained in the constrained reach.

substantial growths of benthic plants. For example, benthic algae and mosses covered 24-29 percent of the active channel in the unconstrained reaches. Constrained reaches, by contrast, were characterized by a single channel with limited sinuosity and poor development of off-channel (lateral) habitat. In the Cascade Mountains, these gorge-like settings are often the result of either active hillslope processes such as earthflows, landslides, and deposition of alluvial fans or passive constraints such as bedrock outcrops that prevent lateral movement of the channel. Both mechanisms tend to funnel flow through a narrow slot on the valley flow. As a consequence, stream gradients are steeper (>3 percent), water depth is greater, and water velocity and stream power are higher than in unconstrained reaches. Due to scour at high flow, sediments tend to be shallow and woody debris accumulation is limited. In the old-growth constrained reach, vegetational diversity and stratification was low because conifers were perched close to the stream channel. Shading of the channel by hillslopes and nearby conifers limited benthic plant cover to only 10 percent of the channel. The second- growth constrained reach was somewhat unique in that it occurred on a broad but inactive alluvial fan. Although the channel was simple and linear, it was wide and its sediments were relatively deep. Light levels on the channel were high because of the absence of large conifers combined with low topographic shading. Accordingly, cover by benthic plants (46 pct.) was the highest of any reach. Retention in Stream Ecosystems Particulate and dissolved matter in transport provides a critical source of nutrients for a substantial portion of the aquatic biota. Particulate organic matter (POM) delivered from adjacent riparian zones into streams includes leaves, needles, seeds, twigs and woody debris. In the stream, this organic matter is decomposed by microheterotrophs such as bacteria and fungi or consumed by detritivorous invertebrates such as insects and snails (Cummins and others 1983). Often, a period of microbial conditioning is necessary before POM becomes palatable to invertebrate consumers (Anderson and Sedell 1979). Dissolved nutrients in streams include inorganic compounds such as nitrogen, often in the form of ionic ammonium (NH+4) and nitrate (NO-3), and phosphorus, as orthophosphate (PO4-3). These nutrients are utilized by stream plants such as benthic algae during photosynthesis, and by certain microheterotrophs such as bacteria during cellular metabolism (Triska and others 1983). Particulate matter is retained in streams primarily by physical trapping. Channel roughness elements, such as coarse sediments and woody debris greatly enhance the retentive efficiencies of stream reaches (Speaker and

Discussion
Geomorphic and Hydraulic Comparisons The unconstrained reaches were characterized by broad valley floors bisected by two stream channels dominated by riffles, rapids, and some small pools. The relatively low gradient (2 percent) and reduced stream power of those reaches allowed accumulation of woody debris in the channel, which further increased channel complexity. Stream sediments were deep and heterogeneous. Lateral habitats were well developed and included secondary channels, backwaters, and isolated pools. Diversity of hydraulic environments was high, and water residence times were great due to considerable subsurface flow. Riparian vegetation was diverse, and included herbaceous, shrub, deciduous, and coniferous species. Stratification of riparian vegetation was apparent, in that herbs and shrubs were located close to the active channel, whereas deciduous and coniferous trees grew at greater distances from the active channel. Because of this stratification, incident sunlight striking the channel was relatively high in unconstrained reaches, which resulted in

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

37

others 1984). Leaves are trapped in the interstices of bed particles and within networks of woody debris accumulations. Along stream margins, low water velocities, shallow depths, and high roughness increase the potential for retention. Thus, complex lateral habitats will enhance leaf retention whereas simple (linear) stream margins will hinder retention. Leaf releases in Lookout Creek indicated that unconstrained reaches were substantially more retentive of particulate matter than were constrained reaches. This retentive efficiency of unconstrained reaches was due to a number of factors, including (1) high roughness, (2) substantial development of lateral habitats (e.g., backwaters and side channels), (3) extensive wetted area, (4) low water velocities, (5) presence of woody debris along stream margins. These factors acted in concert to increase contact between leaves and streambed particles, and thereby enhance trapping of particulate matter. In contrast, constrained reaches were relatively unretentive of leaves. These reaches had lower roughness and wetted area, and thus there was reduced opportunity for contact between leaves and streambed. Poor development of lateral stream habitats in constrained reaches also limited leaf trapping along stream margins.

the old-growth, the unconstrained reach had 2.5 times the nutrient retention of the constrained reach. The unconstrained reach had more algae, higher hydraulic residence, greater water-sediment contact, and deeper sediments than the constrained reach. All of these factors could enhance mechanisms 1- 3 above and thus increase nutrient retention. The unconstrained secondgrowth reach also had substantial nutrient retention, probably because of high rates of algal uptake in this open reach. The alluvial fan that formed the streambed also may have provided the opportunity for ammonium adsorption during subsurface flow.

Summary
Retention of dissolved and particulate matter is of critical importance to the operation of stream ecosystems. Retention largely determines the availability of food resources to aquatic organisms. Long-term retention of detritus delivered from riparian zones, with its subsequent microbial colonization and consumption by detritivores, is critical to energy transfer in most streams. Retention of dissolved nutrients permits levels of primary production and microbial growth necessary to support grazing invertebrates. Higher consumers such as fish that rely on invertebrates for food are thus dependent on retention processes to supply food resources for their invertebrate prey. Retention efficiency is intimately linked to riparian conditions and ultimately to valley floor landforms. Highly complex channels within broad valley floors display high retention, whereas simple channels within narrow valleys are less retentive. Channel geomorphology, including lateral habitat, bedform, and substrate characteristics, is a major determinant of rates of particulate and nutrient retention. Complex strata of riparian vegetation permit light gaps that encourage primary production and thus nutrient retention, while ensuring a steady source of detritus to the stream. Second-growth riparian zones may enhance nutrient retention due to algal uptake, but sacrifice particulate matter retention because of channel simplification. Retention is a complex interaction of valley floor geomorphology, riparian conditions, and in-stream biological demand that accentuates the intimate linkage between aquatic and terrestrial ecosystems.

Dissolved Matter Retention Dissolved nutrients can be retained in stream ecosystems by three different mechanisms: (1) uptake by primary producers, including benthic algae and mosses, (2) absorption or transformation by heterotrophic microorganisms, (3) chemical or physical sorption onto surfaces of inorganic substrates or detritus (adsorption). It is not clear which mechanism(s) was most important for nutrient retention in Lookout Creek. All three processes probably operated to some degree. Mechanism 1 was supported by our observation that retention was highest in reaches with the highest algal abundance, as in the constrained second-growth site. Microbial transformation (mechanism 2) also occurred because as ammonium declined in a downstream direction, nitrate increased slightly, indicating some microbial nitrification (conversion of ammonium to nitrate). Although we have no direct evidence for mechanism 3, as a cation ammonium is readily bound chemically and sorbed physically and thus some removal by this mechanism probably occurred. Regardless of whether the process was uptake, transformation, or sorption, retention serves to conserve nutrients in a specific reach. These nutrients may be used immediately as in the case of uptake or transformation, or stored for later use as in the case of adsorption. High levels of nutrient retention in the unconstrained reaches are probably attributable to the geomorphic and hydraulic complexity of these reaches, which enhances opportunity for nutrient uptake and adsorption. In

Acknowledgments
We thank Kelly Moore for his assistance in data collection and analysis and Gordon Grant and Fred Swanson for providing unpublished data. This research

38

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

was supported by grant number BSR-8508356 from the National Science Foundation.

Newbold, J.D.; Elwood, J.W.; O'Neill, R.V.; VanWinkle, W. 1981. Measuring nutrient spiralling in streams. Can. J. Fish. Aquat. Sci. 38: 860-863. Speaker, R.; Moore, K.; Gregory, S. 1984. Analysis of the process of retention of organic matter in stream ecosystems. Verh. Internat. Verein. Limnol. 22: 18351841. Triska, F.J.; Kennedy, V.C.; Avanzino, R.J.; Reilly, B.N. 1983. Effect of simulated canopy cover on regulation of nitrate uptake and primary production by natural periphyton assemblages. In T.D. Fontaine III and S.M. Bartell, eds. Dynamics of lotic ecosystems. Ann Arbor, Mich.: Ann Arbor Science; 129-159. Vannote, R.L.; Minshall, G.W.; Cummins, K.W.; Sedell, J.R.; Cushing, C.E. 1980. The river continuum concept. Can. J. Fish. Aquat. Sci. 37: 130- 137. Young, S.A.; Kovalak, W.P.; DelSignore, K.A. 1978. Distances travelled by autumn-shed leaves introduced into a woodland stream. Amer. Midl. Nat. 100: 217-220.

References
Anderson, N.H.; Sedell, J.R. 1979. Detritus processing by macroinvertebrates in stream ecosystems. Ann. Rev. Entomol. 24: 351-377. Cummins, K.W.; Sedell, J.R.; Swanson, F.J.; Minshall, G.W.; Fisher, S.G.; Cushing, C.E.; Peterson, R.C.; Vannote, R.L. 1983. Organic matter budgets for stream ecosystems: problems in their evaluation. In J.R. Barnes and G.W. Minshall, eds. Stream ecology. New York: Plenum Press; 299-353. Elwood, J.W.; Newbold, J.D.; O'Neill, R.V.; VanWinkle, W. 1983. Resource spiraling: an operational paradigm for analyzing lotic ecosystems. In T.D. Fontaine III and S.M. Bartell, eds. Dynamics of lotic ecosystems. Ann Arbor, Mich.: Ann Arbor Science; 3-27.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

39

A NEW APPROACH TO FLOOD PROTECTION DESIGN AND RIPARIAN MANAGEMENT1


Philip B. Williams and Mitchell L. Swanson2

Abstract: Conventional engineering methods of flood control design focus narrowly on the efficient conveyance of water, with little regard for environmental resource planning and natural geomorphic processes. Consequently, flood control projects are often environmentally disastrous, expensive to maintain, and even inadequate to control floods. In addition, maintenance programs to improve flood conveyance and enhance levee stability, such as clearing riparian vegetation in channels and on levees, undergo little if any technical scrutiny. Such programs are often prescriptive in nature, rather than based on actual performance standards. A new approach to planning channel modifications for flood damage reduction is presented that is multi-objective and incorporates proper consideration of hydrologic, geomorphic, and biologic factors that influence stream hydraulics.

experienced floods the size of the design flood. Nevertheless, there is now sufficient "operating" experience with artificial flood control channels in smaller floods to now be able to assess the adequacy of some of the conventional hydraulic engineering design criteria, and to propose a new approach to flood control design and riparian management.

Problems with Conventional Flood Control Design


Underestimation of Roughness of Lined Channels Conventional flood control design methodology seeks to minimize the right-of-way required for flood control channels by increasing flow velocities, thereby allowing a narrower channel to be built that reduces flood elevations. This is done typically by lining the channel with smooth reinforced concrete. With a suitable slope in a uniform channel, the low roughness of the concrete can allow "super-critical flow" to develop very fast-moving shallow flow. When super-critical flow occurs, the channel cross-section and right-of-way can be significantly reduced. This has led to many channels of this type being built in California in the last two or three decades. Coastal Northern California has experienced two large floods in the last three decades, in January 1982 and February 1986. The experience of the channelized Branciforte Creek (completed 1959) in Santa Cruz County in 1982, and Corte Madera Creek (completed 1970) in Marin County in 1982 and 1986, shows that these supercritical flow channels do not perform as designed and can overtop their banks at flows considerably smaller than their design flood. Figure 1 shows the channelized Corte Madera Creek overtopping in the 1982 flood when in-channel flows were approximately 4,500 cfs, equivalent to about the 15-year flood. The design flood at this location was 7,800 cfs, with 2 feet of freeboard, equivalent to about the 200-year flood.

Extensive channelization of natural streams has occurred in the last 40 years in urbanizing areas. The purpose of these channelization projects, termed "channel improvements" by hydraulic engineers, was generally to maximize the area and value of developable land by reducing flood hazards. The conventional design methods used for flood control channels were developed mainly based on research carried out in the first half of this century (Brater and King, 1976; Chow 1959). The design methods were based on the application of hydraulics research on sediment-free fluid flow in relatively simple artificial pipe, flume, and weir configurations to modified natural streams. Parallel research in fluvial geomorphology on the behavior of natural streams in flood was not typically incorporated into design methods used by public works engineers. The design methods used today are little different from those used forty years ago. However, with the advent of the computer and hydraulic programs such as HEC-2 (U.S. Army Corps of Engineers, 1982), the analysis of flood elevations and hydraulics is considerably quicker and easier. The usual design standard for flood control projects is to protect surrounding areas against inundation in a 100-year flood or the Standard Project Flood (equivalent to about the once in 200- to 500-year flood). Most channelization projects have only been in existence for 2 or 3 decades, and so very few projects have actually

1 Presented at the California Riparian Systems Conference, September 22-24, 1988, Davis, California. 2 Principal and Senior Associate respectively, Philip Williams and Associates, San Francisco, California.

40

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 Corte Madera Creek during the January 4, 1982 flood overtopping its banks just upstream of a bridge near the College of Marin. Turbulence in foreground is water hitting the bridge face. The rectangular concrete channel is about 2 feet below water surface during a 15-year recurrence flood of 4,500 cfs. The channel was designed to contain a 200-year flood of 7,800 cfs with 2 feet of freeboard. Similar flooding occurred in February of 1986. (Photo: Philip Williams)

On Branciforte Creek during the 1982 flood, the channel filled to the top of its capacity at a peak discharge of 6,650 cfs. This channel was designed to contain a Standard Project Flood of 8,500 cfs with 2 feet of freeboard. The primary reason for the failure of these channels in medium-sized floods was that the actual effective roughness of the channel during the flood was considerably larger than that calculated by conventional design methods. Immediately after the 1982 flood, it was possible to identify the flood profile on Corte Madera Creek by flood marks in the chain-like fence along the channel shown in Figure 1. The best fit analysis using HEC-2 indicates that the actual Manning's roughness at the time of the flood peak

was approximately 0.030 instead of the 0.014 assumed in the design for smooth concrete channel (Vandivere and Williams 1983). The higher roughness meant that flow was "sub-critical" rather than super-critical as designed, and consequently flood elevations were approximately 6 feet higher than predicted for the design flood, even though flows were considerably lower. The primary reason for the increased roughness was the bed load sediment conveyed through the channel. After the flood, boulders up to 12 inches in diameter were observed in a sediment delta formed at the downstream end of the channel. Without considering the effect upon roughness of bed forms, gravel and boulders of this magnitude would be sufficient to cause the increase in roughness that was observed by increasing energy losses at the bed of the channel. A number of researchers in fluvial geomorphology (e.g., Limerinos 1970)

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

41

have demonstrated the relation between bed load size and channel roughness in natural streams. For a Manning's roughness of 0.030, the d84 value would be 2.6 inches (d84 = 84 percent of particles in the sample are smaller than 2.6 inches). In retrospect, the increase in roughness observed in the concrete channel is not surprising. Wherever flood flows are large, large amounts of sediment are mobilized and are conveyed downstream by natural watercourses. This bed material cannot be prevented from entering the concrete channel section, and all of it is conveyed throughout the channel length without the opportunity for deposition in any section. Nevertheless, concrete channels of this type continue to be designed and constructed based on "clear water" analysis, ignoring the direct effects of sediment on the hydraulics. In natural streams, super-critical flow rarely occurs in long reaches due to the size of bed material mobilized (Jarrett 1984). Failure to Account for Channel Bed Erosion and Deposition Clear-water analysis design procedures currently in use do not take into account the significant changes in channel morphology that occur during the course of a flood. Reassertion of meandering in an artificially straightened channel can cause levee failure. However, the most significant effects on flood levels are changes in the channel bed. Hydrologic design criteria for artificial channels assume high antecedent rainfall prior to the design flood event. In California and in many other locations, this creates the conditions for large numbers of debris avalanches and mudflows (National Research Council and U.S. Geological Survey 1984). Large amounts of sediment and debris are introduced into tributary streams and can cause significant aggradation (filling with sediment) of the bed, particularly where the floodplain has been developed, eliminating the natural sediment storage area. This can raise flood elevations and cause flood paths to be substantially different than those predicted. Further downstream, the channel bed can degrade during the course of a flood, lowering flood elevations below those predicted by clear water analysis. An example of the failure to consider the erosion and deposition of sediment in a river channel is the case of the San Lorenzo River Channelization Project in Santa Cruz completed in 1958. The defect in the original hydraulic design, which assumed that the river channel could be maintained at down to about 8 feet below sea level at its mouth, has been documented by Griggs (1984). By 1982, the channel bed had typically silted up about 6 feet. According to clear-water analysis, the

flood control project could only protect against the 30year flood. In January 1982, the flood flow of 30,000 cfs was approximately the 30-year flood, but the flood was contained within the levees. Approximately 4-6 feet of scour had occurred at the downstream end during the time of the flood peak, greatly increasing the capacity of the channel. Current design criteria do not recognize the benefit of keeping a natural sand bed in a channel in reducing flood levels. This leads to construction of flood control projects such as the Los Angeles River flood control channel which have concreted the channel bed, preventing scour in a large flood. Failure to Account for Debris Clear-water design of channels generally ignores or underestimates the role of floating debris in increasing flood elevations. In California and in many other areas, large floods can carry large amounts of debris such as uprooted vegetation and trees, fences, and parts of structures. On smaller streams, this, combined with sediment, invariably impedes or completely blocks the hydraulic efficiency of small- to medium-sized culverts. Further downstream on the main channels, bridges and culverts can be partially obstructed, causing significant increases in flood elevations upstream. On some creeks during large floods, the water surface profile is actually a staircase of obstructed culverts and backwater ponds. The rise in water surface elevation due to backwater from an obstructed bridge can greatly exceed the reduction in water surface elevation due to stream channelization. Figure 2 shows an extreme case a bridge across Soquel Creek in Santa Cruz County obstructed by debris during the 1982 flood (9,700 cfs, or about a 16-year flood event) (Thompson 1982). Flood elevations were increased at least 10 feet upstream and directed the main flood flow out of the channel. Underestimation of Maintenance Requirements The engineering perception that the design of flood control projects is mainly a question of selecting the appropriate channel geometry, has led to an emphasis on initial construction of a flood control project. This in turn has led to neglect of consideration of how flows, sediment, and vegetation interact in determining floodelevations in modified streams. Very little analysis has been carried out of realistic maintenance requirements. Instead, assumptions are made concerning stream geometry and channel roughness, and these are imposed as maintenance requirements on the channel, whether or not they are cost-effective.

42

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Typically, in trapezoidal earth-lined flood control channels, a design Manning's roughness of about 0.03 to 0.04 is used. According to a somewhat subjective interpretation of roughness values for low-gradient natural streams, the design engineer may determine that no woody vegetation can be allowed in the channel bed and banks, and any that grows there must be regularly removed as part of a prescriptive maintenance procedure. Wide trapezoidal channel beds exposed to the sun are ideal nurseries for riparian vegetation such as

willows and cattails, and costs of removal can be high. Frequently, local flood control districts have insufficient money for maintenance and economize by carrying out prescriptive maintenance every few years instead of every year. This can result in the vegetation being managed at a state at which it offers greatest resistance to flows short dense brushy vegetation in the channel bed. Figure 3 illustrates how roughness changes with the age of riparian vegetation.

Figure 2A debris jam of Soquel Creek at the upstream side of the Soquel Drive Bridge after the January 4-5, 1982 flood. This jam (27,000 yd') diverted most of the flood discharge (estimated at 9,700 cfs and about 15-year recurrence) through town (visible in upper left background). Flow depths of up to 6 feet exceeded predicted 100-year elevations. (Photo: Gary B. Griggs)

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

43

A New Approach to Flood Control Design


With an experience of the last few decades, we suggest a design process that will lead to greater longterm reduction in flood damages while allowing the enhancement of riparian corridors. The following are the key elements of this process: 1. Utilize an integrated planning process: This requires an understanding that stream modification will affect more than flood levels. All the significant hydrologic, geomorphic, ecologic, and economic factors have to be considered, rather than approach the design as a plumbing problem. This generally requires involvement of a range of skills beyond traditional hydraulics engineering. 2. Clearly Identify Design Objectives: Stream modifications are rarely single-purpose projects. They typically can include the following: Flood damage reduction (it is important to state this goal in this fashion rather than the nebulous and impossible "control" of floods); Protecting or restoring riparian ecosystems; Figure 3 Conceptual diagram of change in channel roughness with age of riparian vegetation. Other maintenance costs that are frequently underestimated are sediment, debris, and garbage removal and repair of channel bank erosion. The lack of effective maintenance can often negate the reductions in flood elevation initially achieved by channelizing the stream. Providing recreational access; Enhance property values along creek corridor. 3. Understand the physical system: This means developing an understanding of the natural hydrology and geomorphology of the particular watershed and then identifying past, and possible future, human-caused influences on these physical processes. Such an understanding provides the setting in which to establish specific design criteria for a particular reach. 4. Carry out an integrated design: An integrated design would consider not only the direct effects of stream modification on flood elevation but also all the significant processes that affect flood elevations and are affected by the channel modification. Typically, these would include: Downstream effects on hydrology and stream morphology; Effect of future changes in watershed on hydrology and sediment delivery; Relationship between riparian vegetation management and channel hydraulics; Effect on seasonal streamflows; Effect on groundwater levels and recharge; Effect on fisheries; Effect of changes in flow velocities on bank erosion, downcutting, and upstream drainage system;

Current Design Methods Overestimate Channelization Benefits Failure to recognize the design problems described earlier tends to result in an exaggerated expectation of benefits of stream channelization over alternatives that preserve a more natural creek corridor. Recognition of realistic roughnesses and the role of sediment and debris would tend to reduce the supposed benefits of lined channels in favor of preserving flood plains and providing adequate bridge crossings. Recognition of natural scouring during floods would call into question the rationale for lined channel beds. Adequate consideration of maintenance requirements would recognize the hydraulic benefits of more natural riparian vegetation and channel morphology over geometric cross-sections.

44

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Planning for the consequences of failure of any part of the flood management system, e.g., levee failure or culvert obstruction; Designing to minimize long-term maintenance requirements, taking into account the evolution of the stream corridor. 5. Develop maintenance program based on performance standards: The methods used presently for flood control channel maintenance are generally prescriptive in nature. Typically, there is a maintenance program at a set time interval to strip vegetation and regrade channels whether or not it is actually required. Such maintenance practices are not only environmentally destructive but can be expensive and not particularly effective. A performance standard-based maintenance program would establish maximum design floodwater surface elevations. Periodic monitoring of the stream channel, including cross-section surveys and inspection of potential obstructions and channel roughnesses would be required. The results of this monitoring would be used in standard hydraulic programs such as HEC-2 to determine in what portions of the channel sediment removal or vegetation thinning may be required. This approach could significantly reduce the frequency, extent, and environmental disruption of channel maintenance. It would tend to allow riparian vegetation to reach maturity, thereby shading the channel and reducing roughness (see Figure 3). It would also allow the channel the opportunity to flush out some of its accumulated sediments in smallsized floods.

References
Brater, Ernest F.; King, Horace W. 1976. Handbook of hydraulics. 6th ed. New York: McGraw-Hill. Chow, Ven Te. 1959. Open-channel hydraulics. Tokyo: McGraw- Hill Kogakusha Ltd. 680 p. Griggs, Gary B. 1981. Flood control and riparian system destruction: Lower San Lorenzo River, Santa Cruz County, California. In: Warner, Richard E. and Hendrix, Kathleen M., editor, California riparian systems: ecology, conservation, and productive management. Berkeley, University of California Press, p. 142-149. Jarrett, Robert D. 1984. Hydraulics of high-gradient streams. In: Journal of Hydraulic Engineering, 110(11). Limerinos, J.T. 1970. Determination of the Manning coefficient from measured bed roughness in natural channels. USGS Water Supply Paper 1898-B. 47 p. National Research Council and U.S. Geological Survey. 1984. Debris flows, landslides, and floods in the San Francisco Bay Region, January 1982. 1982 Overview and summary of a conference held at Stanford University, August 23-26, 1982. Washington, D.C.: National Academy Press. 83 p. Thompson, W.R. 1982. A flood hazard assessment of the town of Soquel. Unpublished thesis. University of California at Santa Cruz. U.S. Army Corps of Engineers. 1982. HEC-2 water surface profiles: programmers manual. Vandivere, William B; Williams, Philip B. 1983. Flood analysis of the January 4, 1982 storm: Corte Madera Creek, Marin County, California. Unpublished report supplied by author. 21 p. + tables.

An Example of Integrated Design Wildcat Creek


The Wildcat Creek flood control project in North Richmond, California, is an example of the application of many of the aspects of the integrated design approach described above. This project was originally proposed more than 20 years ago as a concrete channel, then as a single-purpose trapezoidal earth channel. Concern by local citizens and environmentalists led to the adoption of a multi-purpose design by the Army Corps of Engineers and Contra Costa County Flood Control District. The key elements of this adopted design, referred to as the "consensus plan," are shown in Figure 4 and are contrasted with the earlier trapezoidal plan. This project, which is intended to reduce flood damages in the adjacent community, restore the riparian corridor, and provide public recreational access, is now under construction.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

45

Figure 4 Conceptual designs for Wildcat Creek Flood Control Project.

46

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

EFFECTS OF BANK REVETMENT ON SACRAMENTO RIVER, CALIFORNIA1


Michael D. Harvey and Chester C. Watson2 Abstract: Twelve low radius of curvature bends, half of which were rivetted, were studied in the Butte Basin reach of Sacramento River, California, to determine whether bank revetment deleteriously affected salmonid habitat. At low discharge (128.6 cubic meters/s) it was demonstrated that revetment does not cause channel narrowing or deepening, nor does it prevent reentrainment of gravels on point bars. Point bar sediments in rivetted bends are not coarser than those in non-rivetted bends. Point bar morphology is stagedependent, and therefore, point bars are both sources and sinks for spawning-size gravel which can mitigate against reduced gravel recruitment due to bank revetment. or lesser than 2.5 (Nansen and Hickin 1986; Harvey 1988). Measured lateral migration rates range from 37 to 10 meters/year (Harvey 1989). Bends are generally rivetted when Rc/W is between 2.3 and 3.8 because the rates of bank erosion are high. The principal concerns on Sacramento River are that revetment will: (1) cause the river to deepen and narrow, (2) cause a coarsening of sediments on the point bars, (3) prevent re-entrainment of gravels deposited on the point bars, and (4) prevent gravel recruitment from the eroding banks. The objective of this investigation was to determine whether revetment of low radius of curvature bends in the Butte Basin reach was adversely affecting the salmonid habitat by modifying the morphologic characteristics of the bends.

Approximately 20 percent of the total bank length (both banks) in the Butte Basin reach of Sacramento River has been rivetted to prevent bank erosion and meander migration that have threatened flood-control levees and flood-relief structures. Revetment in this paper refers to the bank protection method of sloping back an eroding bank (2H:1V) and covering it with rock. A number of concerns have been expressed about the possible effects of bank revetment on the morphologic characteristics of the bends on the river, which in turn may have adverse effects on the salmonid fishery (CA Department of Fish and Game (DFG) 1983; CA Department of Water Resources (DWR) 1984). Approximately 57 percent of the total spawning area in the river is associated with meander bends (fig. 1), and meander bends are also important juvenile rearing areas (DWR 1984). Chinook salmon select spawning areas with a narrow range of physical conditions. Spawning conditions appear to be optimum when the gravel size in the substrate is between 25 and 152 millimeters, flow depth is greater than 0.25 meters, and flow velocity is between 0.2 and 1.5 meters/s (Diebel and Michny 1986). Rearing area: are generally shallow and have low velocity (DFG 1983). Preferred salmon spawning areas are at pool-riffle interchanges (fig. 1) in a bend (Reiser and Bjornn 1979). Bank erosion rates and the rates of lateral migration of the channel are highest on Sacramento River when the bend radius of curvature to width ratio (Rc/W) is 2.5, and they are both lower when Rc/W is greater

Figure 1 Schematic diagram of a meander bend that shows the locations of surveyed cross sections. The cross-hatched areas are preferred salmon spawning areas and the percentages indicate the importance of these with respect to all spawning areas along the Middle Sacramento River (DWR 1984). Few studies have been conducted to determine the effects of bank revetment on river morphology. Friedkin (1945) in a laboratory study of a model stream with sand bed and sand banks observed that revetment caused thalweg deepening and vertical accretion of the point bars in bends with a wide range of radii of curvature, but channel narrowing occurred only in bends with a

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Geomorphologist and Hydraulic Engineer, respectively, Water Engineering & Technology, Inc., Fort Collins, Colorado.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

47

high radius of curvature. The concerns expressed about the effects of revetment (DFG 1983; DWR 1984) appear to be based on the results of this study. Following revetment of a low radius of curvature bend on Red River, Arkansas, water-surface width increased, flow depth decreased and cross-section area did not change. Therefore, revetment did not cause channel narrowing or deepening (Water Engineering and Technology, Inc. 1987). Fall River, Colorado, is a very sinuous, sand and gravel transporting meandering stream with banks that are root reinforced. The extreme resistance of the bank materials is analogous to the effects of bank revetment. Point bars on Fall River are dynamic, stage-dependent geomorphic features even when the concave bank does not retreat (Anthony and Harvey 1987a, 1987b; Harvey and others 1987).

Data Analysis Mean and standard deviations for each of the variables for the revetted and non-revetted bends were determined and the means were tested for statistically significant differences with a t-Test (90% probability level). The results of the t-Tests of the means for all the revetted (18) and non-revetted (18) cross sections (table 1) indicate that: (1) water-surface width in revetted bends is less than that of non-revetted bends, (2) there is no difference in maximum depth, (3) average depth is less in revetted bends, and (4) cross-section area is less in revetted bends. When the non-revetted cross sections were subdivided on the basis of the cohesiveness of the concave bank toe sediments and were compared to the revetted cross sections, the statistical analyses indicate that: (1) there is no difference in water-surface width, (2) the maximum depth is associated with cohesive sediments, (3) average depth is less in revetted bends, and (4) cross-section area is less in revetted bends. These results indicate that the effect of revetment on watersurface width is equivocal, but revetment does cause a reduction in average flow depth and cross-section area. The deepest (maximum depth) flows were associated with the occurrence of cohesive sediments in the toe of the concave bank. Table 1 Comparison of cross section variables at revetted and non-revetted bends of Sacramento River. Discharge was 128.6 cms.
Type of Bend revetted (No. Obs.) Non-revetted (No. Obs.) Non-CohesiveToe (No. Obs.) Cohesive Toe (No. Obs.)
1 2

Sacramento River Field Study

Data Collection Under low-flow conditions (128.6 cubic meters/s) twelve bends in the Butte Basin reach were investigated. Half of the bends were revetted and half were not. The radii of curvature of both sets of bends ranged from 381 to 572 meters. Harvey (1989) demonstrated that the morphology and dynamics of bends on Sacramento River could be related to the radius of curvature of the bends. At each bend three cross sections were surveyed to a common datum with a boat-mounted fathometer and an EDM-theodolite (fig. 1). The cross sections were located consistently at: (1) the upstream limb of the bend (upper), (2) the bend apex (middle), and (3) the downstream limb of the bend (lower). In the non-revetted bends the toe sediments of the concave bank which were primarily composed of point bar sands and gravels (non-cohesive) or abandoned-channel fills (cohesive) were recorded at each cross section. Wolman counts (Wolman 1954) of the lower point bar sediments were made at the head of each point bar to determine the size distribution of the sediments because the sediments at this location are the coarsest on the point bar (Bluck 1971). From the cross-section surveys the following data were obtained: (1) water-surface width, (2) maximum flow depth, (3) average flow depth, and (4) cross-section area below the water surface. Average flow velocities were calculated from the continuity equation (Q - A.V). The survey data was also used to determine the watersurface slope through the bend. Grain-size distribution parameters were obtained from the Wolman counts.

Water Surface Width (m) 106.7+20.5 18 119.5+20.8


1 18

Max. Depth (m) 3.7+1.5 18 4.2+1.6


2 18

Average Depth (m) 1.6+0.8 18 2.3+0.7


1 18

Cross Section Area (m2 ) 166.3+69.7 18 263.4+55.0


1 18

118.6+17.7
2 13

3.7+0.9
2 13

2.1+0.5
1 13

246.1+46.1
1 13

121.6+29.7
25

5.7+2.1
15

2.7+1.0
15

308.2+178.9
15

Significantly different from revetted value at 90% probability level. Not significantly different from revetted value at 90% probability level.

The revetted and non-revetted cross sections were compared on the basis of their locations within a bend: upper, middle, lower (table 2). The results of the statistical tests indicate that: (1) revetment has no effect on water-surface width, nor maximum flow depth, at any of the locations within a bend, (2) revetment causes a reduction in average flow depth in the upper and middle cross sections, but has no effect on the lower cross section within a bend, and (3) revetment causes a reduction in cross-section area in the upper and middle cross sections, but it has no effect on the lower cross section within a

48

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

bend. The results indicate that the effects of revetment are limited to the upper and middle parts of a bend, and the principal effect is a reduction of average flow depth. Water-surface slopes and grain-size parameters for the revetted and non-revetted bends were compared statistically (table 3). The results indicate that revetment has no effect on water-surface slope or on the grain-size distributions of the point bar sediments.
Table 2 Comparison of cross section variables at different
locations in revetted and non-revetted bends of Sacramento River. Discharge was 128.6 cms. Cross Section Upper: revetted Non-revetted (No. Obs.) Middle: revetted Non-revetted (No. Obs.) Lower: revetted Non-revetted (No. Obs.)
1 2Not

Water Surface Width (m)

Maximum Depth (m)

Average Depth (m)

Cross Section Area (m2)

110.6+27.7 127.7+24.7 26

3.0 + 1.3 3.9 + 0.4 26

1.3+0.6 2.2+0.3
16

136.3+47.5 271.0+12.7
16

109.1+19.0 112.5+20.0 26

3.4 + 0.9 4.3 + 1.3 26

1.6+0.7 2.3+0.5
16

164.1+69.4 256.2+32.4
16

99.7+19.0 117.8+17.5 26

4.8 + 1.6 4.5 + 2.5 26

2.1+1.1 2.4+1.2 26

198.4+84.5 262.7+94.5 26

Significantly different at 90% probability level. significantly different at 90% probability level.

reduction in average flow depth in the upper and middle parts of a bend, which may increase the area of rearing habitat in a bend (DFG 1983). The reduction in average flow depth is not sufficient to reduce flow depths below those required for spawning (0.25 meters: Diebel and Michny 1986). The effects of revetment on watersurface width are equivocal because the statistical analyses have produced conflicting results (tables 1 and 2). Revetment does not induce a coarsening of the point bar sediments (table 3), nor does it appear to affect watersurface slope (Friedkin 1945). However, because of continuity, a reduction in cross-section area at a constant discharge must result in higher average velocity in revetted bends. When all of the cross sections (36) were considered, the mean flow velocity + 1 standard deviation of the revetted and non-revetted bends were 0.97 + 0.35 and 0.57 + 0.15 meters/s, respectively. The results of a tTest indicate that the means were different statistically. There was no statistically significant difference between the mean velocities in revetted and non-revetted bends at the lower cross section. However, the mean velocity (0.97 + 0.30 m/s) of the revetted middle cross sections was significantly different from that of the middle non-revetted cross sections (0.56 + 0.08 m/s). This was true also for the upper cross sections (1.16 + 0.39 m/s vs. 0.52 + 0.02 m/s). Therefore, revetment appears to cause an increase in the average flow velocity in the middle and upper parts of the bend. However, the increased velocities are well within the limits of 0.2 to 1.5 meters/s required for spawning (Diebel and Michny 1986). The effects of revetment on gravel re-entrainment from point bars could not be addressed directly in this investigation. However, data from the Butte Basin reach of Sacramento River are available, (Gundlach and Murray 1983), where monumented cross sections were resurveyed at different discharges (fig. 2). The repeat surveys demonstrate that point bar morphology is stagedependent, and that vertical accretion occurs during higher discharges. This accords with the observations of Anthony and Harvey (1987a, 1987b) and Harvey and others (1987). There is no doubt that revetment of a bend prevents recruitment of gravel from the floodplain at that location. Vanoni (1987) concluded that 85 percent of spawning-size gravels are derived from bank and bar erosion, and therefore, revetment will reduce gravel availability. However, to some extent the reduced availability from bank erosion sources is mitigated by the fact that point bars are both sinks for and sources of gravel because they accrete vertically during higher flows and are eroded during recessional flows (Anthony and Harvey 1987a, 1987b; Harvey and others 1987) provided that there is an upstream source of gravels.

Table 3 Comparison of water-surface slopes and grain-size variables at revetted and non-revetted bends of Sacramento River. Discharge was 128.6 cms. Type of Bend revetted (No. Obs.) Non-revetted (No. Obs.)
1ds

Water Surface Slope (m/m) 0.00013+0.000083 26 6

d50 (mm)

d95 (mm)

1ds

(mm)

15.1+6.9 38.6+12.7 2.1 + 0.3 26 26 25 6 6 5

0.000101 +0.000091 12.4+4.7 42.5+14.0 2.1 + 0.4

is geometric standard deviation 2Not significantly different at 90% probability level.

Discussion
This investigation of the effects of revetment on salmonid habitat, which was conducted under low-flow conditions, when the alleged effects of revetment are considered to be most deleterious (DFG 1983), has addressed some of the principal concerns about the effects of revetment on river morphology. Revetment does not cause an increase in channel depth, but in fact appears to cause a

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

49

References
Anthony, D.J.; Harvey, M.D. 1987a. Response of bed topography to increased bedload, Fall River, Colorado. Int. Assoc. Hydrol. Sci. Publ. No. 165; 387-388. Anthony, D.J.; Harvey, M.D. 1987b. Stage-dependent point bar adjustments, Fall River, Colorado. Trans. Amer. Geophys. Union, 68(4); 1297. Bluck, B.J. 1979. Sedimentation in the meandering River Endrick. Scottish Journal of Geology, 7; 93-138. California Department of Fish and Game. 1983. Sacramento River and tributaries bank protection and erosion control investigation: evaluation of impacts on fisheries. State of California, Department of Fish and Game Report; 92 p.

Figure 2 Cross section adjustments with change in discharge in a revetted bend. Data from Gundlach and Murray (1983).

California Department of Water Resources. 1984. Middle Sacramento spawning gravel study. State of California, Department of Water Resources Report, August, 1984; 34 p. Diebel, R.; Michny, F. 1986. Sacramento River, Butte Basin reach salmon spawning study. Report U.S. Fish and Wildlife Service, Division of Ecological Services, Sacramento California, December, 1986; 32 p. Friedkin, J.F. 1945. A laboratory study of the meandering of alluvial rivers. U.S. Army Mississippi River Division, Waterways Experiment Station, Vicksburg, Mississippi; 176 p. Gundlach, D.L.; Murray, L.A. 1983. Data collection program, Sacramento River bend study, California. U.S. Army Corps of Engineers, Sacramento District, March, 1983; 24 p. Harvey, M.D. 1989. Meanderbelt dynamics of the Sacramento River, California. (These proceedings). Harvey, M.D.; Pitlick, J.; Hagans, D.K. 1987. Adjustments of point bar morphology during a snowmelt runoff period. Trans. Amer. Geophys. Union, 68(4); 1297. Nanson, G.C.; Hickin, E.J. 1986. A statistical analysis of bank erosion and channel migration in Western Canada. Bulletin Geological Society of America, 97; 497-504. Reiser, D.W.; Bjornn, T.C. 1979. Habitat requirements for anadromous salmonids. USDA Forest Service, General Technical Report, PNW -96; 54 p. Vanoni, V.A. 1987. Sedimentation aspects of the Sacramento River between Bend Bridge and Colusa. Report to U.S. Army Corps of Engineers, Sacramento District, May, 1987; 29 p. Water Engineering and Technology, Inc. 1987. Geomorphic and hydraulic analysis of Red River from Shreveport, LA to Dennison Dam, TX. Report to U.S. Army Corps of Engineers, Vicksburg District, Contract No. DACW3886-D-0062/7, August 1987; 226 p. Wolman, M.G. 1954. A method of sampling coarse river-bed material. Trans. Amer. Geophys. Union, 35; 951-956.

Conclusions
Revetment of individual bends in the Butte Basin reach of Sacramento River does not effect salmonid habitat adversely. Revetment does not cause channel narrowing or deepening, nor does it prevent re-entrainment of point bar gravels or cause coarsening of the point bar sediments. Gravel recruitment from point bars mitigates to some extent the elimination of the banks as gravel sources provided that there is an upstream source of gravels.

Acknowledgments
We thank Ed Sing of the Sacramento District, Corps of Engineers for his assistance, and Z.B. Begin, Geological Survey of Israel, S.A. Schumm, Colorado State University, and two anonymous reviewers for their constructive reviews. The study was funded by Contract No. DACW05-87-C-0094, U.S. Army Corps of Engineers.

50

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

POST-FIRE INTERACTIONS BETWEEN RIPARIAN VEGETATION AND CHANNEL MORPHOLOGY AND THE IMPLICATIONS FOR STREAM CHANNEL REHABILITATION CHOICES 1
Susan C. Barro, Peter M. Wohlgemuth and Allan G. Campbell
2

A heat wave in southern California during the summer of 1985 set the stage for a record fire year. One of the largest fires began on July 1, 1985 when an arsonist ignited a blaze in the Wheeler Gorge area of the Los Padres National Forest in Ventura County, California. Of the 118,000 acres (47,790 ha) burned in the Wheeler Fire, over 60 percent was scorched by high intensity fires, resulting in severe watershed damage and the production of hydrophobic soils. ore than 30 miles (50 km) of stream channel were also damaged in the fire. The size and intensity of the blaze, as well as its proximity to the pictur-esque town of Ojai, caused managers great concern over the possibility of mudslides and flooding during the first post-fire winter. Plans for emergency rehabilitation, which included reseeding and channel clearing, were implemented immediately. To investigate post-fire interactions between riparian vegetation and channel morphology, we selected a 100 m section of burned riparian zone situated in a 1.5 mi2 (4 km2) portion of the upper Santa Ynez River drainage basin (3429'N, 11926'W)(Figure 1). The sampling scheme involved repeated channel surveys and tree inventories from fall 1985 through winter 1988. Located along a boulder-bed reach about 2.8 mi (4.5 km) upstream of Jameson Lake, a water supply reservoir serving the city of Montecito, the vegetation is typical of a South Coast Riparian Forest: Alnus rhombifolia (white alder) is the dominant species, but Platanus racemosa (sycamore), Populus fremontii(cottonwood), and Quercus agrifolia (coast live oak) are also common. The intensity of the fire through the riparian area is indicated by the damage to trees and the understory. The understory specieswhich included Ribes, Rosa, and Toxicodendronwere completely removed by the fire and only bare ash remained. Over 90 percent of the alders in the studied reach were killed by fire (Figure 2). Many of the oaks, sycamores and cottonwoods were top-killed but were vigorously sprouting from the base. Although there

appeared to be no correlation between tree diameter and mortality, we speculate that bark thickness may affect differential mortality. Fire, in addition to removing vegetation, causes changes in soil characteristics which lead to accelerated erosion on hillside slopes. In fall 1985, the first rainstorms after the fire flushed hillslope debris into the channels, clogging them with sediment. High winds occurring in December 1985 toppled many dead branches and some whole trees into the channel. Subsequent storms organized the downed trees into loosely structured dams of organic debris occurring on average every 10-15 m. These logjams affected the sediment dynamics of subsequent flows, as accumulated sediments were flushed downstream and trapped behind these barriers. The amount of sediment stored behind these structures was quite varied and ranged from one tenth to five cubic meters. Logjams were not noticeable immediately after the burn; rather pre-fire channel configuration was determined by the position of large boulders. Post-fire riparian vegetation/channel interactions can be dictated by management practices. The USDA Forest Service Burned Area Rehabilitation Handbook (FSH 2509.13) suggests that the best management practice balances downstream value protection with the environmental implications of the treatment. To protect life and property, treatments generally involve channel clearing, necessarily at the expense of the riparian environment. Because trees in the riparian zone (both living and dead) act as stabilizing elements in streambed configuration, their removal will provoke adjustments in channel morphology. In addition, removal of obstructions increases flow velocity; which may scour the channel bed, increase the sediment load, degrade the water quality, export nutrients out of the system, and cause deterioration of the biotic habitat.

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California Botanist, Hydrologist, and Hydrologist, Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, Riverside, California

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

51

Channel clearing is often undertaken even when life and property are not at risk. If downstream values include water supply or flood control reservoirs (as along the Santa Ynez River), a more appropriate course of action may be to leave the burned riparian zone intact. While more sensitive to debris flows produced by catastrophic storms, during more moderate events, obstructions will dissipate some of the flow energy, and the in-

creased sediment storage capacity will buffer the sedimentation impacts downstream. Logjams will create a stepped channel gradient, resulting in a slower routing of nutrients, increased water quality, and greater habitat diversity. Although logjams constitute an additional channel configuration parameter, ascertaining the influence of specific obstructions on channel morphology was beyond the scope of this study.

Figure 1Location map of the study area on the upper portion of the Santa Ynez River 52
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

NUMBER OF INDIVIDUALS
Figure 2Tree species frequency and degree of mortality within the study area

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

53

MEANDERBELT DYNAMICS OF THE SACRAMENTO RIVER, CALIFORNIA1


Michael D. Harvey2 Abstract: A 160km-long reach of Sacramento River was studied with the objective of predicting future changes in channel planform and their effects on water-surface elevations. Planform data were used to develop regression relationships between bend radius of curvature (Rc) and both short-term (5 years) and long term (90 years) lateral migration rates (MR) and migration distances (MD). A dimensionless cutoff index (Rc/MD) was developed to predict bend cutoff occurrence. Cutoffs occur when the cutoff index value is between 1.7 and 3.7. Channel planform controls water-surface elevations and bend cutoffs can reduce upstream water-surface elevations by up to 1 meter over a wide range of discharges. and the shape of the bend. The rate of bank retreat is dependent on the resistance to erosion of the concave bank materials (Nanson and Hickin 1986), the duration and magnitude of the flows (Odgaard 1987), the radius of curvature of the bend (Nanson and Hickin 1983, 1986; Odgaard 1987), and the capacity of the flows to transport bed- material sediment (Neill 1984; Nanson and Hickin 1986). Channel migration is a discontinuous process because it is dependent on the occurrence of flood flows (Brice 1977;

In order to obtain an understanding of the meander dynamics of the Sacramento River, which is a coarsegrained meandering river located in the Great Valley of California, a geomorphic study of a 160km-long reach of the river from Colusa to Red Bluff was undertaken in 1987 (fig. 1). The reach of river between Glenn and Chico Landing (Butte Basin), which is located at the upstream end of the flood-control levees, is of major importance to flood control in the lower Sacramento Valley. The objectives of this study were to see if: (1) an understanding of meander dynamics could be used to predict the rates and locations of within-channel erosion and deposition due to changes in river planform, and (2) the planform of the river has significant effects on overbank flooding and sedimentation. Point-bar development and concave bank erosion have been the principal concerns of those studying the dynamics of meandering rivers. Figure 2 is a schematic diagram of a reach of a meandering river that defines the terms that are used in this discussion of the dynamics of the Sacramento River. Erosion along the concave bank occurs because of convective acceleration in downstream flow (Henderson 1966), and because of intensification of cross-stream flow. Both are caused by flow convergence which implies that the shape of a meander bend significantly affects bank erosion (Nanson and Hickin 1983). As the radius of curvature of the bend decreases, the channel cross section in the pool zone is constricted laterally because of vertical growth of the point bar (Knighton 1984; Carson 1986). Therefore, lateral migration of the channel and concave bank erosion are dependent on the flow characteristics
1 2

Figure 1 Location map for study reach of Sacramento River.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Geomorphologist and Hydraulic Engineer, respectively, Water Engineering & Technology, Inc., Fort Collins, Colorado. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

54

Figure 2 Schematic diagram showing in planform the geomorphic surfaces and features that are associated with meander bends. Nanson and Hickin 1983). Initially bends migrate in a cross-valley direction (extension) (fig. 2), but eventually bends advance in the down- valley direction (translation) (Brice 1977; Knighton 1984; Leeder and Bridge 1975; Nanson and Hickin 1983, 1986). Meander bends eventually cut off when the radius of curvature decreases below a certain value which is specific to each stream. Reduction of the radius of curvature of a bend causes backwater upstream of the bend, and this is expressed physically as a reduction in the slope of the water surface. Since the sediment transport capacity of the flows is proportional to the slope of the water surface, a reduction in slope reduces the sediment transport capacity of the flows. This causes deposition of sediment in the upstream limb of the bend between the pool and riffle (fig. 2). Deposition of sediment reduces the flow capacity of the channel and this causes flows to be diverted over the point bar (fig. 2). These flows erode the point bar surface and form chutes (Carson, 1986; Lisle, 1986). However, cutoffs can occur as a result of either chute development (Lewis and Lewin 1983; Brice 1977) or neck closure (Fisk 1947). The review of literature suggests that changes in river behavior should be predictable. Any prediction of future behavior is based on past behavior, streams being no exception (Schumm 1984). Implicit in this approach is the assumption that the sequence of hydrologic events (i.e., flood flows) that have controlled the behavior of the river in the past will be repeated.

Channel Migration and Bank Erosion


Data on channel planform (1:12,000) were available from the California Department of Water Resources (DWR) River Atlas (DWR 1984) and 1986 aerial photographs (1:400). Hydrologic records from 1879 were available at the U.S. Geological Survey gaging station at Bend Bridge, which is located upstream of Red Bluff. The planform data were used to construct a data base which could be used to investigate channel migration and bank erosion in the 32km-long reach between Glenn and Chico Landing (fig. 1). Changes in rivers are generally associated with large floods and, therefore, it is important to differentiate between short-term and long-term behavior. Short-term behavior of the river was based on a 5-year period of record between 1981 and 1986 because large magnitude flow events occurred in both 1983 and 1986. Long-term behavior of the river was based on the period of record (1896 to 1986).

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

55

Short-Term Migration Rates Bagnold (1960), Leeder and Bridge (1975) and Nanson and Hickin (1983, 1986) have demonstrated that lateral migration rates of meandering rivers can be correlated with the radius of curvature (Rc) of bends. Migration rates are highest when the radius of curvature to channel width (W) ratio (Rc/W) is about 2.5. Radii of curvature (Rc) and 1981-1986 migration rates (MR) for 11 bends were measured to obtain short-term data on river behavior. Radii of curvature ranged from 381 to 838 meters and migration rates varied from 37 to 10 meters/year. A least- squares regression of the data is: MR = 53.57 0.049Rc (R2 = 0.69) (1) Figure 3 The ratio of migration rate (MR) to channel width (W) plotted against the ratio of radius of curvature (Rc) to width. The asterisks and bars represent the means and standard errors, respectively. The curves are from Nanson and Hickin (1986).

Long-Term Migration Rates

In order to determine long-term behavior of the river, radii of curvature and migration rates for the period of record (1896- 1986) were utilized. The radii of curvature were assigned to 9 class intervals that varied by 76-meter increments from 229 to 838 meters. The average channel width in each bend was determined by measurement from the DWR Atlas and the 1986 aerial photographs, and both the migration rate and radius of curvature were divided by the channel width (Nanson and Hickin 1986). The average width of the river in the study reach is 150 meters. The relationship between the radius of curvature-width ratio (Rc/W) and the migration ratewidth ratio (MR/W) is shown in figure 3. Also shown in figure 3 are the means and standard errors for the 9 class intervals and the curves which define the upper limits of Nanson and Hickin's (1986) data. The mean migration rates on Sacramento River are close to Nanson and Hickin's maximum values. This suggests that migration rates on Sacramento River are high in comparison with the Canadian sand and gravel streams. The relationships between the mean migration rates and the radii of curvature were established on logarithmically-transformed data. For the radii of curvature greater than 381 meters (Rc/W > 2.5) the least-squares regression is: MR = 6.98 x 104Rc-1.333 (R2 = 0.83) (2)

The reason for subdividing the data is provided by figure 3. Nanson and Hickin's (1986) curves show that for Rc/W values between 1 and 2.5 there is a direct relationship between MR/W and Rc/W. Conversely, for Rc/W values greater than 2.5 there is an inverse relationship between MR/W and Rc/W. Brice (1977) considered that most bends on Sacramento River would cutoff by the time that the radius of curvature had reduced to 381 meters. However, a number of low radius of curvature bends (less than 381 meters) are located in the lower part of the study reach near Colusa. This may be due to the fact that the sediments are finer, more cohesive and, therefore, more resistant to erosion. For a given class interval of radius of curvature between 381 and 838 meters (Rc/W > 2.5) the longterm migration rates are lower than short-term rates by 57 to 73 percent. This is consistent with the general observation that large magnitude, low frequency, events significantly effect changes in the river, and that change occurs because of the occurrence of large floods. Further, both the short-term and long-term data indicate that bends with radii of curvature of about 380 meters erode the fastest, but bends with radii of curvature that are either greater or lesser than 380 meters have lower rates of erosion.

and for the radii of curvature less than 381 meters (Rc/W < 2.5) the least-squares regression is: MR = 2.2 x 10-6Rc2.875 (R2 = 0.94) (3)

56

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Radius of Curvature and Cutoffs


The progressive development of a meander bend to the point where it cuts off is an example of exceeding an intrinsic geomorphic threshold (Schumm 1977). The cumulative frequency distributions of the radii of curvature of bends between Glenn and Chico Landing in 1969, 1981 and 1986 were determined from aerial photographs. The radii of curvature were assigned to 8 equally-spaced class intervals (305 to 838 meters). Between 1969 and 1986 the median radius of curvature declined from 600 to 550 meters. The median radius of curvature for a cutoff is 380 meters, but the range is from 305 to 610 meters. Ninety percent of all cutoffs occur when the radius of curvature is less than 533 meters. The radii of curvature of bends that had cut off since 1908 (10) and pre-1908 meander scars on the floodplain (22) were measured. The radii of curvature of four bends that had cut off following revetment were also measured. Statistical analyses (t-Tests, 90% probability level) of the cutoff data were conducted. The results indicate that there are no statistically significant differences between the mean radii of curvature values of the floodplain (417 + 98m), post- 1908 (419 + 95m) and revetted (390 + 163m) cutoffs. This can be interpreted as indicating that changes in hydrology and upstream sediment supply due to dam construction and gravel mining (DWR, 1984) have had little effect on the meander dynamics in the Butte Basin reach. This may be due to the fact that the dams have not significantly reduced the peak flows and that sediment supply to this reach has been maintained by within-channel erosion in the reach between Red Bluff and Chico Landing (fig. 1). A dimensionless cutoff index, which is defined as the ratio of the radius of curvature to the migration distance (Rc/MD) was developed to predict cutoff occurrence. Equation 1 was used to determine the MD values for the cutoff index for both the recent (10) and floodplain (22) cutoffs. With the exception of two floodplain cutoffs, the Rc/MD values were less than 4. The mean and standard deviation for the recent cutoffs were 2.7 and 1.0, respectively and the values for the floodplain cutoffs were 2.6 and 0.9, respectively. Therefore, cutoffs can be expected to occur when the value of the cutoff index (Rc/MD) lies between 1.7 and 3.7. The cutoff indices for 14 bends between Glenn and Chico Landing were calculated using measured values of MD between 1981 and 1986 (table 1). The data indicate that seven of the bends have Rc/MD values that lie within the range of values that were identified for cutoffs (1.7 < Rc/MD < 3.7). Associated with these Rc/MD values for these seven bends are two other characteristics that were identified on the 1986 aerial photographs: (1)

the presence of a mid-channel bar in the upstream limb of the bend, and (2) the presence of chutes across the point bar. Therefore, it appears that cutoffs can be predicted on the basis of the value of the cutoff index and the presence of the two ancillary features. This was tested on the bend at river distance 278.4km which had cutoff in 1986. This bend was revetted prior to 1981 and, therefore, no migration of the bend took place between 1981 and 1986. However, the radius of curvature of the bend decreased from 572 meters in 1981 to 343 meters in 1986. The MD value for a radius of curvature of 343 meters (Eq. 1) is 181 meters and, therefore, the cutoff index (Rc/MD) is 1.9. The aerial photographs confirm the presence of both a mid-channel bar in the upstream limb of the bend and the chutes on the point bar. Table 1 Characteristics of bends between Glenn and Chico Landing, Sacramento River. Presence of features Radius Short- term River Curvature Migration Cutoff 1986
Distance (Rc) 1986 (km) (m) 307.8 306.4 304.8 304.0 303.0 301.6 299.2 297.6 293.9 292.8 288.0 287.2 285.6 280.0 838 495 381 229 533 572 572 533 838 572 533 572 381 686 Distance (MD)(m) 49 152 186 55 21 155 162 88 61 143 149 116 146 40 Index (Rc/MD) 17.1 3.3 2.1 4.2 25.4 3.7 3.5 6.0 13.7 4.0 3.6 4.9 2.6 17.2 Upstream Bar No Yes Yes Yes No Yes Yes No No Yes Yes No Yes No Revetted Chute Bank No Yes Yes Yes No Yes Yes No No Yes Yes Yes Yes No No No No No Yes Yes Yes No No No No Yes Yes No

Channel Planform and Water-Surface Elevations


The reduction in the radius of curvature of a bend increases the hydraulic resistance of the flow, which causes increased backwater upstream of the bend. This is expressed by a reduction in the slope of the water surface upstream of the bend which reduces the conveyance capacity of the channel and, therefore, promotes overbank flows. In order to demonstrate the effects of channel planform on water-surface elevations, step-backwater runs (HEC-2) were conducted for the reach of river between Gianella Bridge and Woodson Bridge (fig. 1). Gaging stations are located at both bridges, which permits calibration of the water-surface profiles. Discharges of 360,

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

57

850 and 1980 (bankfull) cubic meters/s. were routed through cross sections that were surveyed in 1923 and 1987. The HEC-2-derived water-surface elevations were plotted against a straight-line valley distance (projected profile) upstream of Colusa, which is in contrast to the normal practice of using river distance. Since river distance changes through time, the use of a projected profile allows different surveys to be evaluated on a common basis (Harvey and others 1988). The 1923 and 1987 HEC-2-derived projected watersurface profiles for the three discharges are shown in figure 4. Two bends, located at 66.2 and 76 kilometers above Colusa, cut off between the surveys in 1970 and 1976, respectively. The effects of the planform on upstream water-surface elevations can be seen very clearly in 1923. At the three routed discharges the watersurface elevations increase by about 1 meter upstream of the bends. Following the cutoff at 66.2 km in 1970 the water-surface elevations for the three discharges were reduced by about 1 meter upstream of the bend because the channel downstream was relatively straight and, therefore, did not create any backwater. In contrast, following the cutoff at 76 km in 1976 the water- surface elevation upstream of the bend increased or remained about constant because bends downstream of the cutoff were causing backwater. The comparative water-surface profiles demonstrate that channel planform and its changes have significant effects on water-surface elevations, but the effects of cutoffs also depend on the planform of the channel downstream of the location of the cutoff. Because of the backwater effects of short radius of curvature bends, overbank flooding and sedimentation will occur more frequently in locations upstream of these types of bends. However, a cutoff may or may not reduce the extent of overbank flooding and sedimentation. The overbank effects of a cutoff will depend on the presence or absence of downstream bends.

Figure 4 Projected water-surface profiles for discharges of 360,850 and 1980 cubic meters/s derived from 1923 and 1987 cross sections between Gianella and Woodson Bridges. The solid lines represent the 1923 water-surface profiles and the dashed lines represent the 1987 watersurface profiles. Cutoffs occurred in 1970 at 66.2 km and in 1976 at 76 km. The design-flood capacity of the leveed reach is 4286 cubic meters/s and, therefore, flows in excess of this discharge must overflow through 3 flood-relief structures (FRS) if the integrity of the levees is to be maintained. Satisfactory operation of the flood overflows is dependent on the location of the channel with respect to the flood-relief structures, and on the planform of the river downstream of the structures. Therefore, meander migration and bend cutoffs could have serious consequences for flood control on Sacramento River. Successful operation of two of the FRS is currently threatened by potential changes in the bends at 285.6km (Parrott FRS) and 304.8km (M and T FRS). Bank erosion has been prevented by revetment at 285.6km, but the cutoff index is 2.6 (table 1) and both chutes and a mid-channel bar are present. If this bend cuts off, it is highly likely that the water-surface elevation of flood flows in the loUSDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Application
The ability to predict changes in river planform is important for managing rivers for erosion and flood control. Prediction of future changes is dependent on understanding the past behavior of the river, but uncertainty in prediction is introduced because of the stochastic nature of flood events which cause the changes. On the Sacramento River the ability to predict future changes is especially important in the Butte Basin reach (fig. 1). The Butte Basin reach is a naturally occurring flood overflow area at the head of the leveed section (fig. 1). 58

cation of the Parrott FRS will be reduced by about 1 meter because the downstream reach is straight (fig. 4). Continued bank erosion can be expected at 304.8km (M and T FRS) because it is not revetted. The short-term migration rate is 36 meters/yr (Eq. 1), and the longterm rate is 25 meters/yr (Eq. 2). The cutoff index is 2.1 and chutes and a mid-channel bar are present. A cutoff of this bend will not on its own cause reduced water-surface elevations at the M and T FRS because the channel downstream is sinuous and it is causing backwater upstream (fig. 4). However, continued erosion of the concave bank (fig. 2), and down-valley migration of this bend has the potential to cause a neck cut off at the next bend downstream, and therefore, in the longer term overflows through the M and T FRS are threatened.

References
Bagnold, R.A. 1960. Some aspects of the shape of river meanders. Washington, D.C.: U.S. Geological Survey, Professional Paper 181E; 135-144. Brice, J. 1977. Lateral migration of the middle Sacramento River, California. Sacramento, CA,: U.S. Geological Survey, Water Resources Investigation 77-43; 51 p. Carson, M.A. 1986. Characteristics of high-energy meandering rivers: the Canterbury Plains, New Zealand. Bulletin Geological Society of America 97 (7); 886-895. Department of Water Resources 1984. Middle Sacramento River spawning gravel study. Sacramento, CA.: Calif. Department of Water Resources Report, August, 1984; 38 p. Fisk, H.N. 1947. Fine-grained alluvial deposits and their effect on Mississippi River activity. Vicksburg, MS.: U.S. Army Corps of Engineers Waterways Experiment Station, Report volume 2; 121 p. Harvey, M.D.; Pranger, H.H.; Biedenharn, D.S.; Combs, P. 1988. Morphologic and hydraulic adjustments of Red River from Shreveport, La. to Fulton, Ak., between 1886 and 1980. In: National Conference Proceedings. New York: American Society of Civil Engineers, Hydraulics Division. 1988; 764- 769. Henderson, F.M. 1966. Open channel flow. New York: MacMillan; 522 p. Knighton, D. 1984. Fluvial forms and processes. Baltimore, Maryland. Edward Arnold; 218 p. Leeder, M.R.; Bridge, P.H. 1975. Flow separation in meander bends. Nature 235; 338-339. Lewis, G.W.; Lewin, J. 1983. Alluvial cutoffs in Wales and the Borderlands. International Association of Sedimentologists, Special Publ. No. 6; 145-154. Lisle, T.E. 1986. Stabilization of a gravel channel by large streamside obstructions and bedrock bends, Jacoby Creek, northwestern California. Bulletin Geological Society of America 97; 999-1011. Nanson, G.C.; Hickin, E.J. 1986. A statistical analysis of bank erosion and channel migration in Western Canada. Bulletin Geological Society of America, 97 (8); 497-504. Neill, C.R. 1984. Bank erosion versus bedload transport in a gravel river. In: Proceedings of the River Meandering Conference; 1983, New Orleans, La.; New York: Am. Soc. Civ. Eng.; 204-211. Odgaard, A.J. 1987. Streambank erosion along two rivers in Iowa. Water Resources Research, 23 (7); 1225-1236. Schumm, S.A. 1977. The fluvial system. New York: Wiley-Interscience: 338 p. Schumm, S.A. 1984. River morphology and behavior: problems of extrapolation. In: Proceedings of the River Meandering Conference; 1983, New Orleans, New York: La.; Am. Soc. Civ. Eng.; 16-29.

Conclusions
Lateral channel migration, which involves both point bar deposition and concave bank erosion (fig. 2), can be predicted by meander bend radius of curvature. Longterm rates of channel migration (Eq. 2) vary from 57 to 73 percent of short-term rates (Eq. 1) in bends whose radius of curvature ranges from 381 to 838 meters. Ninety percent of bends cut off when the radius of curvature is less than 533 meters. Bend cutoffs can be expected to occur when the cutoff index values are between 1.7 and 3.7. Meander bend cutoffs can reduce upstream water-surface elevations by up to 1 meter, and therefore, overbank flooding is highly dependent on channel planform changes. The ability to predict changes in river planform and their effects on watersurface elevations provides a rational basis for managing the Sacramento River.

Acknowledgments
I thank Ed Sing of the Sacramento District, U.S. Army Corps of Engineers, for his assistance in obtaining much of the data; Ben Pennock of Glenn-Colusa Irrigation District supplied the 1987 cross-section data for the hydraulic analyses. I thank Z.B. Begin (Geological Survey of Israel) and S.A. Schumm (Colorado State University) and two anonymous reviewers for their constructive reviews.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

59

SESSION B: CENTRAL VALLEY RIPARIAN HABITATS


To the majority of people who responded to the first expressions of concern over the catastrophic loss of riparian habitats in California, the focus was entirely on the magnificent forests of mixed deciduous trees that bordered the rivers of California's Great or Central Valley. The plants primarily at risk were trees. People could see them as forests but the thought was that it was cottonwoods and oaks and the large arboreal willows, along with a scattering of ash and box elders, that were in need of preservation. At the 1981 Riparian Systems Conference, a number of excellent papers described the great biological wealth and historical extent of riparian forests in Valley. In the years that have followed that landmark conference the focus has shifted in three distinct ways. The obvious shift was toward identifying and preserving, through purchase or other means, important examples of these forests. A second shift, arising directly from the emphasis at the 1981 Conference on systems within the riparian environment, was to recognize that there is much more at stake than just trees. Not only is the forest an extraordinarily complex community, but there is a mosaic of habitats involved, including grassland, ponds, varied successional stages leading to several different forest types, and an abundance of associated wetlands. The third shift was toward active management efforts aimed at maintaining ecological functions and enhancing the natural values of those riparian systems. The papers that we present here include one that addresses the question of how much riparian habitat there actually was in the Central Valley historically. This is followed by three papers that deal with protection efforts and surveys for protection purposes. One describes an effort to designate sites for a national registry of natural landmarks. Another seeks to link potential sites with conditions that govern their appropriateness as preserves. A third deals with a recently initiated effort to preserve and restore riparian systems along the San Joaquin River near Fresno, where the concern is dominated by educational values of this scarce habitat. The final two papers are drawn from the poster sessions. They are notable for being oriented toward other valueson the one hand, wildlife populations, and on the other, the linkage between riparian systems that are essentially terrestrial and the adjoining marshlands, now virtually gone. F. Jordan Lang Environmental Scientist Jones & Stokes Associates, Inc. Sacramento, California Dana L. Abell Conference Coordinator University Extension University of California, Davis

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

61

FEASIBILITY OF MAPPING RIPARIAN FORESTS UNDER NATURAL CONDITIONS IN CALIFORNIA1


David R. Dawdy2 Abstract: The California State Water Resources Control Board is conducting hearings to set quantity and quality standards for river flows into San Francisco Bay. Comparisons of present conditions with "natural conditions" prior to European settlement were introduced into the hearings. Consumptive use relations were developed for various riparian and water-related plants, and estimates of the total annual volume of runoff production and consumptive use were developed. A discussion of those computations and a general discussion of the relation of historic riparian forests to "natural flows" into San Francisco Bay is presented. Sacramento River upon the subsidence of each flood, covering the lands flooded for only a period of two or three days. . . From a short distance below Stony Creek to its mouth, the Sacramento River has built up its banks on both sides by the deposit of sediment from the overflowing flood waters, to an elevation of from five to fifteen feet above the basins left on each side." (McClure, 1925, p. 19). McGowan (1961) wrote "valley streams were generally confined to their banks. However, when a river in flood escaped... its banks and spread out over the surrounding countryside, it dropped its coarser sediments first along the edge of the river and... built a natural levee up to ten feet in height on top of its banks... As flood waters moved farther away from the river... their velocity was further decreased so that they now deposited the finer type of sediments... The overflow waters, which built up the land so that it slanted away from the river, eventually reached a trough. . . where they met water being drained from streams which flowed toward them. These troughs, called "basins", became the reservoirs for winter flood waters...Originally, tules, fifteen feet high, covered these basins from which water did not drain off sometimes until August." The Sacramento River overflowed its natural banks when peak flow exceeded stream capacity. The natural levees, basins and overflow lands, and sloughs which drained them created the environment for the riparian system.

The California State Water Resources Control Board presently is holding hearings concerning standards for fresh water inflows into San Francisco Bay. During those hearings, the State Department of Water Resources (DWR) introduced data concerning "unimpaired flows" into the Bay Delta (DWR, 1987). They stated "Unimpaired flow could be synonymous with natural flow if all of the items in the unimpaired computation matched the natural flow computation." The data were "better described as unimpaired data, primarily because of the difficulty in computing four items of significance." One of those four items was "Consumptive use of the riparian vegetation... of the Central Valley under a natural state could be significant but (is) difficult to estimate." A first step in the reconstruction of natural flows would be to estimate the areal extent of riparian forests under natural conditions.

Sacramento River System


The Sacramento River and its major tributaries built natural levees along their banks. When they overflowed, the water slowed in the vicinity of the overbank and deposited the sediment that was carried by the river. The deposition created natural depressions which drained through sloughs. When enough water entered the basin to fill it to overflowing, the water would cut through the natural levee and erode a channel or slough. "The connecting slough channels, together with the steep slope of the trough, provide sufficient drainage so that all flood waters above Chico Creek flow back quickly into the

Travelers' Accounts of Riparian Vegetation


Descriptions of the Central Valley, such as those above and by earlier explorers, can be used to determine the areal extent of the riparian vegetation before extensive modification by European settlement. The routes described by individuals in their diaries can be compared with published maps of vegetation to help to modify or verify those maps. Kuchler (1977) and Fox (1987) have published maps which can be used to plot estimated routes based on written accounts. (Fox's map is used later in this paper because it is at a good scale, about 1:2,000,000, to show most of the Delta and the riparian forest on one page.) Some accounts of areal extent

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Hydrologic Consultant, San Francisco, California (415) 681-0957.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

63

of riparian vegetation can be used with the maps and travellers' accounts. Extent of Natural Tule Areas Cronise wrote in 1868 "Along the San Joaquin River, which spreads out into numerous sloughs, there is, in the northwestern part of (San Joaquin County), an immense expanse of tule marshnot less than 200,000 acres, much of which is covered at all times by a few inches of water, nearly the whole being submerged at high stages of the tide. Late in the season...large sections of these lands become dry on the surfacethe dense body of rushes... having meantime wilted and dried up, the latter often take fire, and burning with terrific fierceness and living tules. In all counties of tule lands, these fires are common, generally occurring in the fall and winter." Based on Cronise's figures and calculations based on his statements, the estimate of the area of tule lands in the Central Valley in 1868 is about 600,000 to 700,000 acres before any major drainage projects. Savannas, Not Tules, Recorded Spanish Expeditions The first Spanish explorers described the immense plains in the Delta. On September 23, 1776, a joint river and land expedition of Spanish explorers started up the river (Bolton, 1926), but only the land exploration continued, and "as soon as it crossed the mountains through a pass of low hills... found itself in the plain which is crossed by the large river (San Joaquin)... they were much farther up than had been agreed upon... decided to continue through the plain up the river. He (followed) the stream for three entire days, traveling rapidly. . . The plain through which that river runs... is as level as the palm of the hand, without any trees except in the bed of the river. It is an immense plain, for he did not see the end of it... After traveling much further on the bank of the river. . . he crossed the great river (and) On the other side of the river he found that the same plain and level land continued. They traveled over it all day." The Moraga expedition of 1808 gave a similar report (Cutter, 1957). He wrote "Today we left (near presentday Livermore) travelling northeast. After about 12 leagues we arrived at the (West Channel of the San Joaquin in its delta area). Having crossed a branch of that river, we spent the night safely... Leaving camp where it was pitched yesterday, this morning...I continued on toward the east, and after about 2 leagues I found the river and I followed it south for about 4 leagues. No ford could be found in this distance... In the afternoon I sent the corporal in a northerly direction in search of

the ford in the river. He found it, but on the opposite side he was confronted by a very large tular and could not continue." (9 Oct 1808) "Today we... moved to the river discovered yesterday...I sent the corporal downstream... He couldn't reach (the mouth of the American River) because of the abundance of tules." (15 Oct 1808) "Today we left for the (Mokelumne River). I went downstream towards the marshes and found nothing good except an immense oak grove." (20 Oct 1808) "Today we followed the (Merced) river downstream, exploring it to its junction with the San Joaquin.. . the low plains of the river are nitrous to within a distance of 2 leagues, more or less, before reaching the San Joaquin.. . There are some beautiful willow groves..."

von Kotzebue The Spanish were interested in the exploration of the San Joaquin inland from their line of missions. They reported extensively on the tules which grew along the San Joaquin in the Delta. Other explorers tended to go up the Sacramento. The Russian, Admiral von Kotzebue, ventured up the Sacramento to the fork with the Feather. He wrote of his trip (von Kotzebue, 1967) "On the 18th of November (1824) the weather was favorable, and we set out... working our way between the islands into the northern portion of the bay... We reached... at a distance of thirty miles from our ship, the common mouth (Carquines Straits) of the two before-mentioned rivers which here fall into the bay... I ascended the highest hillock on the shore... to the north flowed the broad beautiful river formed by the junction of the two, sometimes winding between high, steep rocks, sometimes gliding among smiling meadows, where numerous herds of deer were grazing. In every direction the landscape was charming and luxuriant... we continued our voyage up the stream... We landed... after working only a few miles, and pitched our tents for the night in a pretty meadow... When we had proceeded eighteen miles from our night camp, and twenty-three from the river's mouth, we reached the confluence of the (Sacramento and San Joaquin) . . . Since the river Pescadores (San Joaquin) was already known, I chose the other... after we had ascended it some miles, a violent ... wind forced us ashore; latitude 38 22' (about mile 38) . . . we were obliged to give up for this day. . . pitched our tents in a pleasant meadow . . . I then climbed a hill... and observed that the country to the west swelled into hills of a moderate height, besprinkled with trees growing singly... Between (the mountains) and the river, the country is low, flat, thickly wooded, and crossed by an infinite number of streams, which divide... it into islands... At sunrise... we continued our voyage... The river now took a north-westerly direction.. . The country on the west bank was of moderate height; that on the

64

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

east was low. The power of the current impeded our progress, though our rowers exerted all their strength.. . ...Early the next morning we prepared for our return, and soon quitted these lovely and fertile plains..."

Phelps William Dane Phelps traveled up the Sacramento in 1841 to visit Capt. Sutter, at his fort at Sacramento. Phelps wrote (Busch, 1983) "About a mile inside the Sacramento we passed the mouth of the St. Joachin... we continued on for about 20 miles. All the distance the banks were low and covered with rush flags or Tules as they are called here. At 11 PM having passed all the Tule, we ran along the high banks on which were many high trees... stopped by the side of a beautiful wood, and dined under the wide spreading limbs of a gigantic oak... The immense size of the trees, the dense thickness of the unpenetrated forests in some places, and the level plains with here and there a bunch of scrub oaks without underbrush in others, together with a profusion of wild flowers... The forests consist mostly of sycamore, a variety of oak, but mostly the white, ash & some wallnut." (29 July 1841) "The river here is broader than below, and the banks higher. The country looks fertile and with just trees enough to give it a pleasant appearance. . . (We) started (for Sutter's Fort) at 1/2 past 4 and riding over a beautiful gently undulating country abounding with rich feed and agreeably diversified with trees & wild shrubbery." M'Collum M'Collum (Morgan, 1960) describes the San Joaquin in 1849. "Bound for the southern mines... we were... going up. . . the San Joaquin to Stockton... The passage up the San Joaquin was a dreary one. The river for the greater portion of the way winds like a tape worm, through low marshy ground, where the tules, grow to an enormous height, not allowing us to see out... As you approach Stockton, the uplands, oak-openings and glades of timber, begin to approach the river. . . The tide of the ocean... sets up here, from one to two feet."

Leonard Zenas Leonard traveled to Suisun Bay from the Merced River in early November 1833. He wrote (Ewers, 1959) "This plain (going down the Merced to the San Joaquin) is well watered and is quite productive, as we found a large quantity of wild pumpkins and wild oats... The land is generally smooth and level, and the plains or prairies are very extensive, stretching...as far as the eye can reach... (rivers run) parallel with each other through the plain... with their banks handsomely adorned with flourishing timber of different kinds... This grove of timber may be found along the river at any point, and generally extends about four miles into the plain. Between this grove of timber and the forest extending from the foot of the mountain, there is a level prairie of the richest soil, producing grass in abundance large prairie covered with wild oats... This plain lays on the south side of the river." (February 15, 1834) "Continued our journey up Sulphur River (San Joaquin), passing through a fine country, most of which is prairie, covered plentifully with wild oats and grass."

Belcher The English Captain Belcher explored the Sacramento in 1837. He wrote (Pierce and Winslow, 1969) "Having entered the Sacramento... The marshy land now gave way to firm ground, preserving its level in a most remarkable manner, succeeded by banks well wooded with oak, planes, ash, willow, chestnut, walnut, poplar, and brushwood. Wild grapes in great abundance overhung the lower trees, clustering to the river...On the 30th (October 1837), about four p.m., we found the deep boats stopped at (the forks of the Feather and the Sacramento)... Throughout the whole extent... the country is one immense flat...Our course lay between banks, varying from twenty to thirty feet above the riverlevel... These were, for the most part, belted with willow, ash, oak, or plane,... which latter, of immense size, overhung the stream... Within, and at the verge of the banks, oaks of immense size were plentiful. These appeared to form a band on each side, about three hundred yards in depth, and within (on the immense park-like extent, which we generally explored when landing for positions) they were to be seen in clumps... wandering over what might otherwise be described as one level plain or sea of grass..."

Bryant Edwin Bryant came overland to California (Bryant, 1985). On August 30, 1846, his party reached Johnson's farm on Bear Creek, 40 miles upvalley from Sacramento. August 31 they "marched south seven miles, and encamped on the bank of a chain of small ponds of water. The grass around the ponds was rank and green, and we were protected from the hot rays of the afternoon sun by the shade of evergreen oaks." They traveled on south down the valley 26 miles to the American River. "The valley of the Sacramento, as far as we have travelled down it, is from 30 to 40 miles in width, from the foot of the low benches of the Sierra Nevada, to the elevated range of hills on the western side. The composition of the soil is trodden up by immense herds of cattle and horses which grazed here early in the spring, when it was

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

65

wet and apparently miry. We passed through large evergreen oak groves, some of them miles in width." They marched down the overflow basin from the Bear River on the east side of the Sacramento to the American River, and noted evidence of previous spring overflows, but no tules. Bryant continued "September 13th.-We commenced to-day our journey from New Helvetia to San Francisco ... we travelled in a south course over a flat plain... and encamped on a small lake. . . near the Coscumne river.. . The stream is small, but the bottom lands are extensive and rich... The grass on the upland plain over which we have travelled, is brown and crisp from the annual drought. In the low bottom it is still green." "September 14.-We crossed the Coscumne river... and travelled over a level plain covered with luxuriant grass and timbered with the evergreen oak, until. . . we crossed the Mickelemes river... and encamped on its southern bank in a beautiful grove of live-oaks... The soil of the bottom... produces the finest qualities of grasses. The grass on the upland is also abundant... Our route has continued over a flat plain, generally covered with luxuriant grass, wild oats, and a variety of sparkling flowers... Large tracts of the land are evidently subject to annual inundations. About noon we reached a small lake surrounded by tule... Passing through large tracts of tule we reached the San Joaquin at dark, and encamped on the eastern bank... The ford of the San Joaquin is about forty or fifty miles from its mouth... Oak and small willows are the principal growth of wood skirting the river... Entering upon the broad plain we passed, in about three miles, a small (alkali) lake. The grass is brown and crisp... We passed during the afternoon several tule marshes, with which the plain of the San Joaquin is dotted. At a distance, the tule of these marshes presents the appearance of immense fields of ripened corn. The marshes are now nearly dry, and to shorten our journey we crossed several of them without difficulty. A month earlier, this would not have been practicable... While pursuing our journey (across the Delta to the Ranch of Dr. Marsh) we frequently saw large droves of wild horses and elk grazing quietly upon the plain." Bryant visited with Dr. Marsh. "Sept.17... After breakfast I walked with Dr. Marsh to the summit of a conical hill... from which the view of the plain on the north, south, and east, and the more broken and mountainous country on the west, is very extensive... The hills and the plain are ornamented with the evergreen oak, sometimes in clumps or groves, at others standing solitary... the San Joaquin, at a distance of about ten miles, is belted by a dense forest of oak, sycamore, and smaller timber and shrubbery. The herds of cattle are scattered over the plain,some of them grazing upon the brown, but nutritious grass; others sheltering them66

selves from the sun, under the wide-spreading branches of the oaks." The southern part of the San Joaquin Delta as seen from its western edge, prior to the coming of the Americans, seemed to have scattered tracts of tules in some parts. Bryant travelled up the Sacramento. "October 24 (1846)... About two o'clock, P.M., we entered the mouth of the Sacramento. The Sacramento and San Joaquin rivers empty into the Bay... at the same point ... and by numerous mouths, or sloughs... These sloughs wander through an immense timbered swamp.. . The banks of the river and several large islands which we passed during the day, are timbered with sycamore, oak, and a variety of smaller trees and shrubbery. . . The islands of the Sacramento are all low, and subject to overflow in the spring of the year. The soil of the river bottom, including the islands, is covered with rank vegetation, a certain evidence of its fertility..." Bryant continued "October 25... As we ascended the stream the banks became more elevated, the country on both sides opening into vast savannas, dotted occasionally with parks of evergreen oak."

Comparison with the Fox Natural Vegetation Map


Figure 1 is a partial copy of Fox's map (1987) of natural vegetation for the middle part of the Central Valley. Point A is in the vicinity where the 1776 Spanish expedition hit the San Joaquin and traveled upstream without mention of trees. The band of riparian forest shown by Fox is 5 miles in width. B1 to B2 show the reach that the Moraga expedition of 1808 explored along the San Joaquin and at the mouth of the Mokelumne and the Merced Rivers. Point C is the farthest point reached by Captain Belcher and by von Kotzebue. They reported oaks and savannah after the first few miles up the Sacramento. Tules shown by Fox north of the Sacramento should be riparian forest and savannah. Point D1 is about the point at which Phelps in 1841 reported he ran along high banks on which were many high trees. This report agrees with those of Belcher and von Kotzebue. Point D2 is where Phelps traveled overland. The area of tules shown by Fox should be much reduced in size. The line marked E is the approximate route of Edwin Bryant in 1846, and is close to the "road across the plains" on Lt. Derby's map of 1849 (Thompson, 1961), which shows the road about 4 miles from the river, with riparian forest a mile wide on the west bank and about three miles wide on the east bank from the Feather to the American.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 Partial copy of Fox's map for Central Valley showing location and routes of early travelers.

Points F1 to F4 trace the route of Bryant. At F1 on the crossing of the Cosumnes he reported grass in the "low bottom" of the river, at F2 on the crossing of the Mokelumne "the bottom... produces the finest qualities of grasses." The line marked F3 is in the vicinity of where Bryant broke off the usual road and struck off across the south part of the Delta. Much of the area shown as tules by Fox was mixed plains dotted with tule marshes which went dry late in the summer. Point F4 is where Bryant climbed a hill south of Antioch and looked over the south Delta.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Estimates of Extent of Riparian Forests


Smith (1977) says "The riparian woodlands occurred on the natural levees formed by the Sacramento, Lower Feather, American, and other aggrading streams... Based on historical accounts, it has been estimated that there were about 775,000 acres of riparian woodlands in 1848-1850." Roberts and others (1977) estimated 800,000 acres in 1848, and Fox estimated 938,000 acres 67

(1987, Table 3) for the Sacramento Basin. Katibah (1984) gave a more detailed listing of the locations of the riparian forest stands in presettlement times. Katibah's listing cannot be matched precisely with Fox's, but an approximate comparison is as follows:
Location Sacramento Basin Delta San Joaquin Basin Tulare Basin Total Fox Table 3 938,000 198,000 298,000 515,000 1,949,000 Katibah Table 1 acres 583,400 100,700 187,500 50,000 921,600

Cutter, Donald C., ed., The Diary of Ensign Gabriel Moraga's Expedition of Discovery in the Sacramento Valley 1808, Glen Dawson, 1957. Ewers, John C., editor, Adventures of Zenas Leonard Fur Trader, Univ. of Oklahoma Press, Norman, 1959. Fox, Phyllis, Freshwater Inflow to San Francisco Bay under Natural Conditions, Appendix 2 to State Water Contractors Exhibit 262, Calif. State Water Resources Hearings on the Bay-Delta, Dec. 1987. Katibah, E. F., A Brief History of Riparian Forests in the Central Valley of California, in California Riparian Forests, R. E. Warner and K. M. Hendrix, editors, Univ. of Calif. Press, Berkeley, 1984. Kotzebue, Otto von, A New Voyage Round the World in the Years 1823- 1826, Henry Colburn and Richard Bentley, London, 1830, reprinted by Da Capo Press, New York, 1967. Kuchler, A. W., The Map of the Natural Vegetation of California, Appendix to Terrestrial Vegetation of California, M. G. Barbour and Jack Major ed., Wiley Publ., New York, 1977. McClure, W. F., Sacramento Flood Control Project Revised Plans, submitted by State Engineer to the Reclamation Board, February 10, 1925, California State Printing Office, Sacramento. McGowan, J. A., History of the Sacramento Valley, Volume I, Lewis Historical Publ. Co., New York, 1961. Morgan, Dale L., California As I Saw It, Pencillings by the Way of its Gold and Gold Diggers! and Incidents of Travel by Land and Water by William M'Collum, M.D. A returned Adventurer, The Talisman Press, Los Gatos, Calif., 1960. Pierce, R. A., and Winslow, J. H., editors, H.M.S. Sulphur at California, 1837 and 1839. Being the accounts of Midshipman Francis Guillemard Simpkinson and Captain Edward Belcher, The Book Club of California, San Francisco, 1969. Roberts, W. G., Howe, J. G., and Major, Jack, A Survey of Riparian Forest Flora and Fauna in California, in Riparian Forests in California, Anne Sands, editor, Univ. of California at Davis, 1977. Scott, L. B., and Marquiss, S. K., An Historical Overview of the Sacramento River, in California Riparian Forests, R. E. Warner and K. M. Hendrix, editors, Univ. of Calif. Press, Berkeley, 1984. Smith, Felix, A Short Review of the Status of Riparian Forests in California, in Riparian Forests in California, Anne Sands, editor, Univ. of California at Davis, 1977. Thompson, Kenneth, Riparian Forests of the Sacramento Valley, California, Annals Assoc. Amer. Geographers, Vol. 51, 1961.

Katibah's estimates are conservative and subject to refinement. Sacramento plus Delta, 684,100 acres is compared with 775,000 of Smith and 800,000 of Roberts. Katibah appears to be about 15 percent below the others. Scott and Marquiss (1984) agree with the 800,000 acres, quoting an unpublished report. Fox appears high.

Conclusions
A more extensive study of the routes of the various travelers, the study of unpublished diaries, and a comparison with published maps and accounts of extent of riparian vegetation can be used to map more thoroughly the areas covered by various types of vegetation during the time immediately prior to European settlement and, in particular, before the time of American settlement in the Central Valley. A more accurate mapping of the riparian forests can be helpful in the determination of natural water use and water availability in the Central Valley.

References
Bolton, H. E., editor, Historical Memoirs of New California by Fray Francisco Palou, O. F. M., Vol. IV, Univ. California Press, Berkeley, 1926. Bryant, Edwin, What I Saw in California, Univ. of Nebraska Press, Lincoln, 1985; reprint of 1848 edition. Busch, B. C., editor, Alta California 1840-1842, The Journal and Observations of William Dane Phelps, Arthur H. Clark Co., Glendale, Calif., 1983. California Department of Water Resources, California Central Valley Unimpaired Flow Data, Second Edition, February 1987. Cronise, T. F., The Natural Wealth of California, H. H. Bancroft & Co., San Francisco, 1868.

68

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

GREAT VALLEY RIPARIAN HABITATS AND THE NATIONAL REGISTRY OF NATURAL LANDMARKS1
Robert F. Holland and Cynthia L. Roye2 Abstract: The National Registry of Natural Landmarks is a program established by the National Park Service that seeks to recognize nationally significant examples of the Nation's natural history. Nearly 100 Great Valley riparian sites were evaluated using Park Service criteria. Three sites illustrative of the range of this biotic theme were recommended to the National Park Service for designation as National Natural Landmarks. ficult for evaluators to locate similar sites to make the required comparisons. In the early 1980's the Western Region of the National Park Service began using a more systematic approach to the identification of potential Landmark sites: 1) regional themes were described: where necessary to encompass the diversity present within the natural region, smaller, recognizable assemblages of plants and animals which repeatedly occur within the theme were identified as subthemes to make site comparisons more feasible; 2) all known good sites of each subtheme were identified; 3) uniform criteria of national significance were applied to each site; and 4) site-specific reports were written for sites best meeting these criteria. The purpose of this study was to research, classify, and describe the Great Valley Riparian Forest theme (including subthemes) of Fenneman's South Pacific Border Region, and to recommend sites illustrative of that theme as National Natural Landmarks. The study was part of a multi-state pilot project in which the Western Region of the National Park Service attempted to use state Natural Heritage Inventories as a source of background information and staff expertise in locating potential Landmark sites. These inventories originally were established by The Nature Conservancy to collect, manage, and use biological, ecological, and related information for conservation planning in an attempt to preserve the diversity of our natural heritage. The Natural Diversity Data Base, which is now part of the California Department of Fish and Game and maintains the California natural heritage inventory, was contracted to prepare a Landmark evaluation of the Great Valley Riparian Forest theme. Landmark status is open to lands within the United States, Puerto Rico, the Virgin Islands, and the Pacific Trust Territories. Lands managed by the National Park Service are ineligible for Landmark recognition. To be considered for designation, a site must meet a set of criteria demonstrating that it has national significance. These criteria were used in evaluating all sites and in comparing their relative suitability for Landmark designation. They are summarized in tabular form.

The National Natural Landmark program recognizes nationally significant examples of ecological or geological features of the Nation's natural history. Since 1962 over 500 Landmarks have been designated by the Secretary of the Interior and listed by the National Park Service in the National Registry of Natural Landmarks. The Landmark program is intended to encourage preservation of sites which illustrate the geological and ecological character of the United States, to enhance the educational and scientific value of these sites, to strengthen appreciation of natural history, and to foster public interest and concern for the conservation of the Nation's natural heritage (Federal Register 1975). The ultimate goal of this program is to ensure preservation of the vast majority of ecosystem types present in this country prior to the activities of European man so as to present an intact picture of our country in its pristine condition (Stebbins and Taylor 1973). Until the 1980's, the primary method used to identify potential Landmark sites was through a series of studies of natural phenomena, or "themes", in each of Fenneman's (1928) 33 physiographic regions of the United States. Both biotic and geologic theme studies were completed. Each study resulted in a description of the major features of a region and an initial inventory of sites considered to be excellent examples of the themes found within the region. Sites were then evaluated in the field according to study-specific criteria and compared to other sites representing the same theme to determine which were most appropriate for designation as National Natural Landmarks. Although many excellent Landmark sites were identified through the regional theme studies, in some cases themes were too broad to allow meaningful comparison of sites. In other cases it was dif-

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Vegetation Ecologist and Research Assistant in Ecology, respectively, Natural Diversity Data Base, The Nature Conservancy and California Department of Fish and Game, Sacramento, CA.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

69

Phase I, Classification
Stebbins and Taylor's (1973) theme study treats all Great Valley riparian habitats as one "Special Ecosystem" which they called Riparian Woodland. Several classifications of such habitats have appeared since Stebbins and Taylor's theme study. These indicated a much broader range of habitat types than any one site possibly could possess. A classification by Jensen and Holstein (1983) was chosen as most appropriate for this study. They recognized three riparian forest units: Great Valley Cottonwood Riparian Forest, Great Valley Mixed Riparian Forest, and Great Valley Valley Oak Riparian Forest. These communities constitute the subthemes of this Landmark study. There are surprisingly few data on the composition and structure of these forest types. These data, mostly in Sands (1977) and Warner and Hendrix (1985), and field work throughout the Great Valley indicated that these three subthemes could be generally described as follows: Great Valley Cottonwood Riparian Forest Subtheme This community of medium to tall (to 100 ft), broadleaved winter-deciduous trees typically has well closed canopies that are densely stocked with Fremont cottonwood (Populus fremontii Wats.) and valley willow (Salix gooddingii Ball.). Understories of various other willow species are common. This subtheme encompasses early seral riparian forests in which the tree canopy has been closed long enough to inhibit establishment of sun-loving species but not long enough for shade-tolerant species to grow into the canopy. Disturbance by high flows occurs most years and appears to be an important factor in stand regeneration. Most stands are even aged, reflecting episodic reproduction of the shade-intolerant dominants. Great Valley Mixed Riparian Forest Subtheme Great Valley Mixed Riparian Forest also is a community of medium to tall (to 100 ft), broad-leaved winterdeciduous trees. The tree canopy is fairly well-closed and is composed of several species including Fremont cottonwood, sycamore (Platanus racemosa Nutt.), California black walnut (Juglans hindsii Jeps.), Goodding's willow (Salix gooddingii var. variabilis Ball.), red willow (Salix laevigata Bebb.), yellow willow (Salix lasiandra Benth.),

and box elder (Acer negundo var. californicum Torrey and Gray). These and other shade-tolerant shrub species such as Oregon ash (Fraxinus latifolia Benth.), and buttonbush (Cephalanthus occidentalis L.) form a dense understory. Lianas such as wild grape (Vitis californicus Benth.), virgin's bower (Clematis ligusticifolia Nutt. in T. & G.), and poison oak (Toxicodendron diversilobum Torrey and Gray) are conspicuous, giving the community an appearance popularly associated with tropical jungle. Great Valley Mixed Riparian Forests usually occupy sites that are farther from the active river channel or that are at somewhat higher elevations relative to the active channel; flooding on these higher sites is less frequent and less intense. Most stands are uneven-aged, reflecting the shade-tolerance of the dominant species and the longer recurrence intervals between major disturbances such as flooding, windthrow, or fire.

Great Valley Valley Oak Riparian Forest Subtheme Great Valley Valley Oak Riparian Forest once occurred extensively along the highest parts of the floodplains where flood damage was least likely and least severe. Where it survives, this forest is dominated by the winter-deciduous valley oak (Quercus lobata Nee.). Canopies are moderately to densely closed and up to 100 feet tall. Individuals of sycamore and Oregon ash frequently are scattered in the canopy and sub canopy. Many valley oak stands have basal area distributions suggestive of episodic reproduction. Shrub canopies are best developed in light gaps caused by fire, flood, or windthrow. Wild grape is the only conspicuous liana, often smothering shrubby patches before being shaded out by the closing tree canopy. Canopy structure typically is less complex than that found in Great Valley Mixed Riparian Forest.

Phase II, Site Evaluation


The Natural Diversity Data Base inventory of Natural Communities provided a preliminary list of about 100 sites for the three subthemes studied. Sites were visited and field survey forms were filled out for each site with reference to the Park Service criteria in Table 1. In some cases the forests no longer existed because of type conversion to agriculture or urbanization. Sites which did not fit Park Service criteria for obvious reasons such as these were eliminated from further consideration as potential Landmarks.

70

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 1. Criteria for evaluating the national significance of proposed National Natural Landmarks Primary Criteria Illustrative Character The site must be representative of the subtheme under study. The site has the species composition and physical features typical of that ecosystem. The site offers an exceptional opportunity to illustrate or interpret the natural history of the nation. Present Condition The site approximates an undisturbed natural environment. The site is essentially free from disturbances which detract from its natural character. Secondary criteria In comparing sites which have met these primary criteria a second set of criteria is used: Diversity In addition to its primary natural feature, the site contains high quality examples of other ecological and/or geological features. The presence of other native plant communities such as marshes, willow thickets, and grassy savannas, in addition to the forest community gave sites an advantage in meeting this criterion. Rarity The site provides high quality habitat for one or more rare, threatened, or endangered species, or contains a rare geological or ecological feature. Value for science and education The site is associated with a significant scientific discovery or concept, has a long-term history of on-site research, or offers unusual opportunities for interpretation and public education about the natural history of the particular subtheme.

For the 53 sites remaining, the evaluator interviewed as many people as possible who had knowledge of the site to gather information about land use history and condition. Contacts included farmers working in adjacent fields, ranchers, local landowners and managers, Department of Fish and Game unit biologists, Soil Conservation Service personnel, irrigation and levee district staff members, and local conservationists. From the site visits combined with these interviews, it proved possible to narrow the field to six to ten sites for each subtheme. This subset of sites was re-evaluated using the same Park Service criteria and, based on this improved knowledge, three sites emerged for recommendation as National Natural Landmarks: South Fork Kern River, Kern County: Great Valley Cottonwood Riparian Forest Subtheme Feather River, Sutter and Yuba Counties: Great Valley Mixed Riparian Forest Subtheme Cosumnes River, Sacramento County: Great Valley Valley Oak Riparian Forest Subtheme.

site, any damage expected to occur from continuation of current management practices, and the effects of publicity should the site be designated as a National Natural Landmark. We provided maps of the site boundaries and land ownership details for each parcel within the proposed Landmark. We evaluated each site's features in relation to the Park Service criteria and offered an opinion, based on our professional judgment, of whether or not the sites merited designation as National Natural Landmarks. We were able to find sites so representative of each subtheme that three were recommended for designation. Management guidelines were prepared that detailed actions needed to ensure that significant features would not be damaged, altered by unnatural means, or destroyed. These suggestions did not carry the implication that the National Park Service would actually manage the site, but rather were suggestions the Park Service could provide to landowners to promote conservation of the area's integrity. We found that parts of all three areas already were under partial management for conservation of their natural values. Existing management plans were instructive for at least a portion of each area. Guidelines for those portions not currently such management were added. South Fork Kern River The South Fork Kern River, Kern County, supports the largest surviving example of the Great Valley Cottonwood Forest subtheme. The site is dominated by

Phase III, Evaluation Report


The next phase of the recommendation process was to write a detailed evaluation of each of these three sites. Evaluations included location, boundaries, and size of the site as well as a thorough description of site's plant communities and animal habitats. We addressed the land use and present condition of the

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

71

Fremont's cottonwood, yellow willow, red willow, and valley willow with a dense understory of saplings of these species, abundant stinging nettle (Urtica holosericea Nutt.), and mulefat (Baccharis viminea DC.). There are several small marshes along natural sloughs and some stands of grassland on high ground within the forest. A number of sensitive birds are known to nest in this forest including about 25 percent of the California Yellowbilled Cuckoo (Coccycus americanus occidentalis Ridgway) population. We recommended 1856 hectares (4583 acres) along the South Fork for Landmark status designation. In addition to the largest surviving cottonwood forest, this site has a diversity of seral stages and a natural hydrograph which is unparalled among the other surviving Cottonwood Forest sites. The site is owned by private parties, The Nature Conservancy, and the United States Army Corps of Engineers. Both the Conservancy and the Corps manage their holdings (about half of the acreage) for their natural values. Most of the privately owned parcels are used for cattle grazing. Feather River The Bobelaine-Lake O'Connor area along the lower Feather River supports the largest known surviving example of the Great Valley Mixed Riparian Forest subtheme. We recommended that 1502 hectares (3756 acres) be designated as Landmarks. Fremont cottonwood and sycamore dominate the diverse forest here, and the varied understory includes box elder, Oregon ash, California black walnut, and willows. The shrub layer is dense and diverse. Draperies of several liana species are impenetrable in some areas. Small stands of the Great Valley Valley Oak Riparian Forest and extensive willow thickets also are found within the recommended Landmark. Several rare riparian-dependent animals use the area for roosting and breeding. These include the Federally listed Endangered Valley Elderberry Longhorn Beetle (Desmerocerous californicus dimorphous Fisher), the California yellow-billed cuckoo, and Swainson's hawk (Buteo swainsonii Bonaparte). Both the cuckoo and the hawk are listed as Endangered under California law. There also is a large rookery of great blue herons (Ardea herodias L.) and great egrets (Casmerodious albus L.), and a thriving population of river otters (Lutra canadensis Schreber) and ringtail cats (Bassariscus astutus Lichenstein). This area is of such high quality that most of it has already been acquired by the California Department of Fish and Game or the National Audubon Society for protection of natural values. Most privately owned parcels are used for cattle grazing or are idle.

Cosumnes River Area This area consists of three tracts along the lower Cosumnes River near its confluence with the Mokelumne. We recommended expansion of a Landmark which was designated in 1976. Taken together, the three tracts recommended by this study are the largest known surviving example of the Great Valley Valley Oak Riparian Forest Subtheme (432 hectares, 1079 acres). All three sites are dominated by healthy, vigorously reproducing stands of valley oak. The sub canopy consists of young valley oaks, Oregon ash, buttonwillow, and box elder. Wild grape lianas are common. The tracts differ in species associated with the dominant valley oak reflecting their differing distances from the river channel. Some small sloughs and quiet backwaters support examples of freshwater marshes and small stands of the Great Valley Mixed Riparian Forest subtheme. Swainson's hawk (State-listed Threatened) and both lesser and greater sandhill cranes (Grus canadensis canadensis L. and Grus canadensis tabida Peters, also State- listed as Threatened) have significant roosting areas here. Most of the area is privately owned and receives summer and fall cattle grazing; a small part is owned by The Nature Conservancy and forms the Cosumnes River Preserve. The Conservancy continues to expand its preserve as additional properties become available. Registry The evaluation report and all comments from land owners, managers and any other interested parties are reviewed by the National Park Service to determine which sites qualify for nomination to the Secretary of the Interior. If the Secretary agrees with the findings of the Park Service, the site is designated as a National Natural Landmark. The National Park Service contacts each landowner to explain the Landmark Program and to invite the landowner to register his or her land as a National Natural Landmark. Registration constitutes a voluntary, non-binding agreement between the Secretary of the Interior and the landowner to preserve the significant natural features of the site. The landowner receives a certificate and bronze plaque indicating the site is a registered National Natural Landmark. The landowner does not relinquish any rights to the land and the agreement may be terminated by either party. The National Natural Landmark Program was created administratively within the National Park Service. No legislation or administrative procedures afford specific protection to Landmarks. Official recognition is the only direct protection afforded. Landowners often find pride in owning a site recognized as being of national significance and may be less inclined to convert the land

72

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

to uses which would deprive the site of Landmark status (The Nature Conservancy 1975). When sites are owned by public agencies, Landmark recognition may sway management decisions toward preservation of natural values and stimulate Cooperative Management Agreements between the agency holding title and agencies mandated to preserve natural resources, such as the California Department of Fish and Game. For areas already under management for the preservation of natural values, such as portions of the three sites recommended by this study, Landmark designation commends the owners for their careful stewardship and increases public awareness. Such designation may enhance the availability of funds for acquisition when and if landowners within the recommended Landmark boundaries chose to sell their property. One further indirect protection is pointed out by The Nature Conservancy (1975). Under the National Environmental Policy Act of 1969, Federal agencies undertaking major Federal actions must file statements detailing the effect of such actions on the environment. If proposed actions will have impact upon a National Natural Landmark, any adequate Environmental Impact Statement should note that fact and actions may be modified as a result. In addition to instituting a more systematic approach, the National Park Service determined that the methods used in this pilot study were less costly, site by site, than Landmark identification and evaluation techniques that had been used in the past. Although not yet in use nationwide, the methods used in this pilot study are now accepted by the Park Service as a valid way to identify and evaluate sites for possible Landmark designation. The project benefited the Natural Diversity Data Base by bringing to light several riparian stands which had not been included in the inventory previously and by providing more recent and complete information for all sites visited.

Farmen, United States Army Corps of Engineers; Monte Knudsen, United States Fish and Wildlife Service; and Darrel Coldani, lessee of a portion of the Cosumnes site, for contributing valuable information about individual sites. We also thank Deborah Jensen, formerly with the California Department of Fish and Game, Gene Wehunt, Western Regional Office, National Park Service and Curt Soper, Western Regional Office, The Nature Conservancy for initiating, organizing, and administering the multi-state effort of which this study was a part.

References

Federal Register, Vol. 40, No. 87, Monday, May 5, 1975 at p. 19504. Fenneman, N.M. 1928. Physiographic divisions of the United States. Assoc. Am. Geogr. Annals 18(4):261-353. Holland, R. 1986. National Natural Landmark evaluation phases I, II, and III. Prepared for the National Park Service, United States Department of the Interior. Sacramento, Calif.: California Natural Diversity Data Base, Nongame-Heritage Program, California Department of Fish and Game. Jensen, D.B.; Holstein, G. 1983. Checklist of California natural communities. Sacramento, Calif.: California Department of Fish and Game, Sacramento, CA. 10 p. Unpubl. ms. Sands, A., ed. 1977. Riparian forests in California: their ecology and conservation. Davis: University of California, Institute of Ecology Publication No 15. 122p. Stebbins, G.L.; Taylor, D. W. 1973. A survey of the natural history of the South Pacific Border Region, California. Davis: University of California. Institute of Ecology Publication No. 4. 513p. The Nature Conservancy. 1975. Preserving our natural heritage. Volume I. Federal activities. Prepared for the United States Department of the Interior National Park Service. Washington, D. C.: Superintendent of Documents, U.S. Government Printing Office. 323p. Warner, R. E.; Hendrix, K. M. 1984. California Riparian Systems: Ecology, Conservation, and Productive Management. Berkeley, Calif.: University of California Press. 1035p. Wehunt, G. 1988. Western Regional Office, National Park Service, San Francisco, Calif. [Telephone interview with C.L.Roye, June 23, 1988].

Acknowledgments
We thank Scott Clemons, Preston Johns, and Dale Whitmore, California Department of Fish and Game; William Dillinger, National Audubon Society; Rick Hewitt and Steve Johnson, The Nature Conservancy; Tony

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

73

PLANT COMMUNITY DEVELOPMENT, SITE QUALITY ANALYSIS AND RIVER DYNAMICS IN THE DESIGN OF RIPARIAN PRESERVES ON THE MIDDLE SACRAMENTO RIVER, CALIFORNIA1
Niall F. McCarten2 Abstract: Loss of riparian habitat along the Middle Sacramento River, over the last 100 years, has reduced a once contiguous riparian forest to a series of disjunct remnants of varying size and quality. With limited financial resources to purchase and protect some of the remaining riparian plant communities, it has become necessary to develop methods to select which of the remaining habitats are to become protected. A site evaluation method was developed that included vegetation quality, type and rarity, size, viability, unique features, rare plants, shape, and potential for growth. Using this method 240 plant community sites that included 5 plant community types were evaluated and ranked. The evaluation method and the results are outlined and a critique of the method discussed in light of long term riparian nature preserve design. ous studies have not, however, compared all the riparian sites as to plant community type, determined the quality of these sites, their size, and ranked them to identify specific areas for preservation. Current riparian preservation strategies for the middle Sacramento River have focused primarily on areas that support rare species. Only rarely have sites been preserved based on vegetation or other features (see Katibah and others 1984). It also appears that the preservation strategies have not considered the long term viability of the sites chosen. The dynamics of the Sacramento River are sufficiently well documented to suggest that many of the largest riparian forests along the river today will likely not persist beyond 20-75 years from now (Scott and Marquis 1984). Further, the normal successional sequence of riparian forests prevents us from assuming that the type of community preserved today will be the same type 20 or more years later (Strahan 1984).

The loss of riparian vegetation and habitats in California has resulted in the need for an immediate coordinated effort by government agencies and private organizations to develop a plan to preserve what little remains. However, the cost of preservation, in conjunction with the unwillingness of some riparian land owners to sell, restricts the amounts and locations of land that can be acquired. In order to develop a strategy for the preservation of the remaining riparian habitats in California three pieces of information are needed: 1) location and types of remaining riparian vegetation, 2) the quality of individual contiguous riparian areas and 3) a preserve plan for which areas should be acquired. The middle Sacramento River, between Keswick Dam and Verona, is a riparian corridor 167 kilometers long. The combined impacts from agriculture, levee building and bank stabilization (rip-rap) has reduced this once contiguous riparian gallery forest to a series of disjunct riparian islands. Previous studies have made a qualitative determination of the types of plant communities that remain along some sections of the middle Sacramento River (Conard and others 1977; Katibah and others 1984; Roberts and others 1977; Sands 1977; Thompson 1961; Warner and Hendrix 1984). Only one previous study by Michny (1984) quantitatively determined that approximately 4,000 hectares of riparian "woodland" have potential for habitat preservation. Previ-

Study Objectives
This paper has two purposes: 1) discuss a method to evaluate and rank individual riparian communities, and 2) discuss the selection of individual riparian sites for preserves considering the dynamics of the Sacramento River. The reason for developing the evaluation method was to provide a riparian community site quality analysis and ranking for the Sacramento River Riparian Atlas (1988).

Methods
The site quality evaluation method developed for the middle Sacramento River study was designed to account for all levels of site quality and size excluding active agricultural land. The study required a vegetation quality analysis for individual plant communities.

Presented at the Symposium on California Riparian Systems; September 22-24, 1988; Davis, California. Graduate Student, Department of Botany, University of California, Berkeley. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

74

Site Variables Eight variables were used to evaluate overall site quality: 1) vegetation quality, 2) community rank, 3) size of community, 4) viability, 5) unique features, 6) rare plants, 7) shape of community stand, and 8) potential for stand expansion. Vegetation Quality Five plant communities occur in the study area that have been classified in the California Department of Fish and Game's natural community system (Holland 1986). These communities are: great valley mixed riparian forest, great valley cottonwood riparian forest, great valley valley oak riparian forest, great valley willow scrub and coastal and valley freshwater marsh. A description of the three riparian forest communities is included elsewhere in these proceedings (Holland and Roye 1989). Vegetation quality was primarily evaluated based on the biological features of the community structure that were determined from field surveys. The four features considered were: 1) tree density i.e. canopy closure, 2) species diversity 3) extent of seedling establishment, mainly for areas of disturbance now undergoing regeneration, and 4) maturity of the community. Vegetation quality was evaluated on a scale of 1 through 4 (i.e, 1=poor, 2=fair, 3=good, 4=excellent). Community Rank Vegetation quality was weighted using a plant community ranking value that reflects the level of rarity for each community. The California Natural Diversity Data Base (CNDDB) has ranked some of the rarer plant communities in California including the riparian forest and the freshwater marsh. The valley oak riparian forest is ranked highest by CNDDB due to its rarity. The cottonwood and mixed riparian forest and freshwater marsh are all ranked equally and lower than the valley oak riparian forest. The willow scrub community is not considered rare and has not been ranked. Community ranking values used were 1 through 3, with 3 representing the rarest community type. These community rankings were multiplied by the vegetation quality values. Size of Community Community size is important in preserve design since it is related to the capacity of an area to buffer external threats (Lovejoy and others 1986). The stand sizes used in this study were for individual plant communities. Four size classes were used and point values assigned as follows: 1 =< 50 acres, 1.5 = 50-75 acres, 2 = 76-100 acres, 3 => 100 acres.

Site Viability The viability of a site is its potential to survive natural or human-caused forces over time. This variable mainly represents the survivorship potential from river realignments. Impacts from river erosion was based on the level of river bank cutting. Viability values ranged from 0 through 5, with a high level of viability equal to 5. Unique Features Sites were assigned points for a variety of features that mainly correspond with overall site diversity. For example ox bow areas that had an associated freshwater marsh were given 3 points, mixed riparian communities that included valley oak as a subdominant were given 2 points. The highest number of points given in this category was 5 for an individual site, but technically the category had no a priori maximum value. Rare Plants Although no rare plants were found during the study, 5 points would have been given for their presence. This category includes state or federally listed species, as well as non-listed rare plants. The point assignment was determined relative to other categories where the value of 5 also was the maximum value. Shape of Community Stand The shape of a preserve has been recognized as an important factor in determining the effects of external disturbances or threats (Janzen 1986, Lovejoy and others 1986). In particular, so-called "edge effects" from neighboring areas having contrasting land use can be important potential threats (Harris 1984, Janzen 1986). Two major areas of potential edge effects for the Sacramento River are river course changes on one side and human-caused disturbances on the other. Site quality of a particular plant community with respect to shape was determined using a circumference-to-area ratio index (CAI). The assumption is that areas having a small CAI have the least amount of exposed vegetation to external threats such as circular shaped areas. In contrast narrow strips of vegetation have a high level of external exposure (high CAI) and a higher potential external threat. Values based on shape range from 1-5 with 5 representing a perfect circle.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

75

Table 1- Riparian sites and total area in each quality level by plant community, middle Sacramento River, California, 1988. Quality Riparian Community Cottonwood Mixed Riparian Willow Scrub Freshwater Marsh Totals Excellent No. Sites 10 6 2 10 50 Good Fair Poor No. Sites No. Sites No. Sites 15 8 6 69 26 7 10 82 12 3 2 33 Total Area(ha) 1,320 614 352 121 5,907

Potential for Stand Expansion Site evaluation of potential revegetation areas was determined independently for the Sacramento River Atlas project (Kraemer 1988). However, some assessment was included to evaluate potential expansion of a riparian community stand. This determination was based entirely on the amount of neighboring habitat. Values ranged from 1 through 5 were added for potential growth as follows: 1 point was given for potential expansion less than doubling the current size; 2 points for doubling the current size; 3 points for tripling the current size etc.

good, fair and poor. Each of these levels were given a corresponding letter (i.e., A = excellent, B = good, C = fair, and D = poor). The reason for the grade-type ranking was due to the requirements of the Sacramento River Riparian Atlas project. The Sacramento River Riparian Atlas mapped each riparian community using a letter grade code. In addition, the program's philosophy was only to consider excellent and good quality sites in choosing areas as potential preserves.

Results
A total of 240 community sites were evaluated including 123 mixed riparian forest, 62 cottonwood riparian forest, 24 valley oak riparian forest, 21 willow scrub, and 10 freshwater marsh community types (Table 1). The total area for all riparian community stands is 5,907 hectares (Table 1). Only 23 percent of the site are off excellent quality, 29 percent of the sites are good quality, 34 percent of the sites are fair quality, and 14 percent of the sites are poor quality. Only 12 percent (720 hectares) of the 5,907 hectares of remaining riparian vegetation on the middle Sacramento River is publicly owned. Approximately half of the publicly owned riparian vegetation can be considered protected. Most of the remaining riparian habitat along the Sacramento River is subject to deforestation which currently is at a rate of about 170 hectares per year (Department of Water Resources 1979).

Site Values Two hundred forty riparian community sites that included the 5 plant community types were evaluated and ranked. Each plant community site was given a site value based on the following relationship: Qs =Vq x R x Si + Vi + U + Pr + Sh + Pg where Qs = community site quality, Vq = vegetation quality, R = community rank, Si = community size, Vi = viability, U = unique features, Pr = rare plants, Sh = community stant shape and Pg = potential for stand expansion.

Site Ranking Each of the 240 site values were used to develop a distribution bar graph for each of the plant communities. Coastal and valley freshwater marsh which had only ten sites were ranked equally. The site value distribution curves often had distinct breaks in value groupings making it relatively easy to assign a four category quality rank. The four quality levels were excellent,

Discussion
This study has provided a detailed analysis of the extent and quality of the riparian plant communities

76

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

that were found along the middle Sacramento River in 1987. The results of the analysis found that it was relatively easy to determine site quality classes using a numerical method. The large number of sites and the range of site quality was the key to this study. The the large variation in site quality produced a good distribution from which to segregate quality classes. Site quality characterization for large numbers of sites, such as in this study, should consider a numeral evaluation method such as this. However, comparison of only a few sites (i.e., < 25) to determine the quality could more easily be done without a numerical analysis. The method presented here was used in the Sacramento River Riparian Atlas (McCarten and Patterson 1988). The vegetation information in conjunction with wildlife habitat analysis, spawning gravel data, and potential revegetation site data have provided an assessment of the location of "high quality" riparian habitats. Similarly, The Nature Conservancy has utilized the vegetation quality data for their riparian preserve program (Phelps 1988). For The Nature Conservancy program Mr. Tod Wells and I added the community site values for all contiguous communities to produce a single cummulative value. The site vegetation values for contiguous riparian areas were then compared with other factors to determine areas for potential riparian preserves (Phelps 1989). The vegetation quality analysis method used here has proved useful for choosing riparian sites that are high quality and under the current philosophy for preserve design help target particular locations for preserves. The results of this study provide the basis on which to develop a long term preserve plan for the Sacramento River. That plan should not only consider the larger more mature sites (i.e. those ranked as "excellent"), from each plant community type. Nor should the focus of land acquisition only be toward rare species. Harris (1984, p. 158) has outlined criteria for selecting specific "habitat islands" as follows: 1) geographic position within the system, 2) intrinsic diversity, 3) particular species, e.g., rare species, 4) contribution to within species genetic diversity, 5) presence of endemics, and 6) contribution to the system of "habitat islands." In essence, the selection of a particular site should be considered on the basis of its overall contribution to the system being preserved. The successional ecology of the riparian plant communities, in response to river course and substrate depositional changes over time, requires a dynamic view of a Sacramento River riparian preserve. Due to the island nature of the remaining riparian sites in conjunction with the dynamics of the system, the concept of "long-rotation islands" (Harris 1984, p. 155) may be useful in preserve planning. This concept has been applied to old growth forest systems (Harris 1984), but

aptly applies to the formation, successional changes and eventual loss of riparian habitats at particular sites.

Recommendations
When considering plant community sites that have been ranked as excellent, good and even fair for preserves other factors must be taken into account. Selected sites should include those that have potential for stand expansion; are buffered from river course changes i.e., ox bow islands; have broad contiguous stretches along the river; have both young and mature vegetation i.e., early and late successional stages such as willow scrub and mixed riparian forest; have rare plants and animals; represent examples from all plant communities that are well distributed along the river; and have incipient depositional areas where the river course is moving away from the new habitat. These recommendations promote the view that the goal of a riparian preserve system should consider more than location, size and quality of the present habitats as they appear. The preserve design should also consider the location, size and quality of habitat years from now. Large riparian gallery forests preserved today are likely to be located in the middle of the Sacramento River bed within the next fifty years. On the other hand small patches of willow scrub and developing cottonwood riparian forest will be the gallery mixed riparian forests in the future.

References
California Department of Fish and Game 1988. Sacramento River Riparian Atlas. Department of Fish and Game, Sacramento, California. 79 p. Conard, S.G.; Macdonald, R.M.; Holland, R.F. 1977. Riparian vegetation and flora of the Sacramento Valley. In: Riparian Forests in California, A. Sands, ed. 47-57. Institute of Ecology Pub. 15. 122p. University of California, Davis. California Department of Water Resources (DWR). 1979. Land use changes in the Sacramento River riparian zone, Redding to Colusa, 1972-1977. No. District Report, June 1979. Available from the Dept. of Water Resources, Sacramento. Harris, L.D. 1984. The Fragmented Forest. Island Biogeography Theory and the Preservation of Biotic Diversity. University of Chicago Press, Chicago. 211 p. Holland, R.F. 1986. Preliminary Descriptions of the Terrestrial Natural Communities of California. NongameHeritage Program, California Department of Fish and Game. Unpublished draft supplied by the author; 156 p.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

77

Holland, R.F.; Roye, C.L. 1989. Great valley riparian habitats and the National registry of natural landmarks. In these proceedings. Janzen, D.H. 1986. The eternal external threat. In: Soul, M.E. ed. Conservation Biology, the science of scarcity and diversity. Sunderland: Sinauer Associates, Inc.; 584 p. Katibah, E.F.; Dummer, K.J.; Nedeff, N.E. 1984. Current condition of riparian resources in the Central Valley of California. In: R. Warner and K. Hendrix, eds. California Riparian Systems. Berkeley, Calif.: University of California Press; 314-322. Kraemer, T.J. 1988. Potential riparian revegetation sites and priorities. In: California Department of Fish and Game 1988. Sacramento River Riparian Atlas. Department of Fish and Game, Sacramento, California. 79 p.

of Engineers. 36 pp. Available from the Army Corps of Engineers, Sacramento, Calif. Phelps, S. 1989. The Nature Conservancy riparian preservation program on the Sacramento River. These proceedings. Roberts, W.G., Howe, J.G.; Major, J. 1977. A survey of riparian forest flora and fauna in California. In: A. Sands, ed. Riparian Forests in California; Davis, Calif.: University of Calif. Institute of Ecology; 3-19. Sands, A. (ed.) 1977. Riparian forests in California: Their ecology and conservation. In: A. Sands, ed. Riparian Forests in California; Davis, Calif.: University of Calif. Institute of Ecology; Publ. No. 15. 122 p. Scott, L.B.; Marquiss, S.K. 1984. An historical overview of the Sacramento River. In: R. Warner and K. Hendrix, eds. California Riparian Systems, Berkeley, Calif.: University of California Press.; 51-57. Strahan, J. 1984. Regeneration of riparian forests of the central valley. In: R. Warner and K. Hendrix, eds. California Riparian Systems, Berkeley, Calif.: University of California Press.; 58-67. Thompson, K. 1961. Riparian forests of the Sacramento Valley, California. Annals of the Association of American Geographers 51:294-315. Warner, R.E. 1984. Structural, floristic, and condition inventory of Central Valley Riparian Systems. In: R. Warner and K. Hendrix, eds. California Riparian Systems, Berkeley, Calif.: University of California Press.; 356-374. Warner, R.E., and K.M. Hendrix. (eds.) 1984. California riparian systems: Ecology, conservation, and productive management. Berkeley, California: University of California Press; 1035 p.

Lovejoy, T.E.; Bierregaard, R.O.; Rylands, A.B.; Malcolm, J.R.; Quintela, C.E.; Harper, L.H.; Brown, K.S.; Powell, A.H.; Powell, G.V.N.; Schubart, H.O.R.; Hays, M.B. (1986). Edge and other effects of isolation on Amazon forest fragments. In M.E. Soul ed. Conservation Biology, The Science of scarcity and diversity. Sinauer Associates, Inc. Sunderland. 584 p. McCarten, N.F.; Patterson, C.A. 1987. Vegetation quality and rare plant study of riparian plant communities along the middle Sacramento River. A technical report submitted to the Nongame-Heritage Program, California Department of Fish and Game. 41 p. Available from the Nongame-Heritage Program, Sacramento, Calif. McCarten, N.F.; Patterson, C.A. 1988. Riparian community sites rare plants and priorities. In Sacramento River Riparian Atlas. California Department of Fish and Game, Sacramento California. Michny, F.J. 1984. Riparian vegetation protection program: An appraisal level study. Report to the U.S. Army Corps

78

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

SAN JOAQUIN RIVER RIPARIAN HABITAT BELOW FRIANT DAM: PRESERVATION AND RESTORATION 1
Donn Furman2 Abstract: Riparian habitat along California's San Joaquin River in the 25 miles between Friant Darn and Freeway 99 occurs on approximately 6 percent of its historic range. It is threatened directly and indirectly by increased urban encroachment such as residential housing, certain recreational uses, sand and gravel extraction, aquiculture, and road construction. The San Joaquin River Committee was formed in 1985 to advocate preservation and restoration of riparian habitat. The Committee works with local school districts to facilitate use of riverbottom riparian forest areas for outdoor environmental education. We recently formed a land trust called the San Joaquin River Parkway and Conservation Trust to preserve land through acquisition in fee and negotiation of conservation easements. Opportunities for increasing riverbottom riparian habitat are presented by lands from which sand and gravel have been extracted.

Table
corridor

1 Riparian wildlife/vegetation
Corridor Acres 1,088 588 400 1,844 101 148 309 606 2,846 Corridor Percent 14.0 7.0 5.0 23.0 1.5 2.0 4.0 7.5 36.0

Category Water Trees Shrubs Other riparianl Sensitive Biotic2 Agriculture Recreation Sand and gravel Riparian buffer
1

Total 7,900 100.0 Land supporting riparian-type vegetation. In most cases this land has been mined for sand and gravel, and is comprised of gravel ponds. Range of a Threatened or Endangered plant or animal species.

Study Area
Natural vegetation in the San Joaquin River bottomlands between Friant Darn and Freeway 99 creates a broad corridor of different vegetation communities. The most critical of these vegetative communities is the riparian community which interfaces water, vegetation and wildlife resources, bisecting or bordering on diverse habitat. Along the 25 miles of the San Joaquin River bottomlands below Friant Darn only about 6 percent of the original riparian habitat may be found (Counties of Madera and Fresno, and City of Fresno 1986). Losses of riparian habitat along the San Joaquin River are attributable to agricultural and urban encroachment, sand and gravel extraction, road construction, snagging, clearing, and riprapping. In 1986 the California Department of Fish and Game mapped the riparian vegetation corridor along the San Joaquin River for Madera and Fresno Counties, and the City of Fresno as they studied the San Joaquin Riverbottom between Friant Dam and Freeway 99 (Counties of Madera and Fresno, and City of Fresno 1986). Figure 1 shows the San Joaquin River riparian vegetation corridor. Table 1 summarizes the results of that mapping by acreage and by category.

The majority of the undisturbed riparian habitat lies between Friant Dam and Highway 41 beyond the city limits of Fresno. Of the 588 acres of riparian woodlands in the San Joaquin River corridor, 264 acres are classified as Great Valley Riparian Forest and lie on and adjacent to the Ball Ranch, approximately 3 miles below Friant Dam (Atlantis Scientific 1987). Discharges into the San Joaquin River are controlled by the Bureau of Reclamation at Friant Dam. Approximately 95 percent of the average annual runoff of the San Joaquin River is diverted at Friant Dam for export south to Kern County and north to Madera County. The Bureau releases enough water into the river channel to deliver 5 cubic-feet-per-second past the last riparian water right holder at Gravelly Ford, about 37 miles downstream. Since instream flows were reduced in 1954, agriculture, urban uses, and sand and gravel extraction have occurred within the 25- to 100- year floodplain between Highway 41 and Freeway 99. Below Highway 41 down to Freeway 99 much of the original riparian woodland has been removed for sand and gravel extraction, for golf courses, and for agriculture. The river in many areas has been diverted to flow through gravel ponds. This portion of the riverbottom also contains many older gravel ponds surrounded by woody vegetation comprised of shrubs, willows, and cottonwoods.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Executive Director, San Joaquin River Committee, Fresno, California.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

79

Figure 1- San Joaquin River riparian vegetation corridor. Mapping of San Joaquin River riparian vegetation corridor along the San Joaquin River between Friant Darn and Freeway 99. Prepared by the California Department of Fish and Game for the San Joaquin River Area Reconnaissance Study. 80
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Threats to San Joaquin River Riparian Habitat Between Friant Dam and Freeway 99
San Joaquin River riparian habitat is threatened primarily by urban encroachment into the floodplain. Two proposed residential developments would place approximately 2,300 units on 835 acres within the river corridor. The proposed Ball Ranch residential development called for removing all but 22 of the 264 acres of Great Valley Mixed Riparian Forest in order to regrade the property to elevate housing pads one-foot above the 100-year floodplain. Residential development directly threatens riverbottom riparian habitat in the San Joaquin River corridor through removal. It also indirectly threatens riparian habitat by increasing the need for future flood control measures such as channelization, clearing, snagging, and riprapping. At the present time the San Joaquin River below Friant Dam is not controlled by levees. Some minor levee work has been done along short stretches to protect sand and gravel processing plants. Local agencies that exercise jurisdiction along the San Joaquin River corridor between Friant Dam and Freeway 99 include Fresno County, Madera County, and the City of Fresno. The City of Fresno placed a moratorium on residential development in 1986. It is currently considering adoption of riverbottom open space zoning that will preclude permanent residential structures. Land uses permitted under this zoning include agriculture, sand and gravel extraction, compatible commercial such as catfish farms, and public and private recreation including golf courses. The City has proposed policies to protect riparian vegetation. The Madera County Board of Supervisors has placed a moratorium on residential development of riverbottom lands in its jurisdiction pending development of a specific plan. The current Madera County General Plan designates all areas not subject to intensive development for continued agricultural, grazing, and open space uses (County of Madera). Fresno County's General Plan calls for open space land uses within their portion of the San Joaquin Riverbottom (County of Fresno 1976). Fresno County's jurisdiction is the largest of the three agencies. It contains most of the remaining undisturbed riparian habitat. The County adopted River Influence Area Policies in 1976, one of the objectives is which to conserve and enhance the natural wildlife habitat. The Fresno County Board of Supervisors have indicated a willingness to consider amending the General Plan to permit residential housing. To date, no developUSDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

ment proposal has been presented to the Fresno County Board of Supervisors for a vote. Another threat to riparian habitat in the San Joaquin Riverbottom is potential recreational development. A 20-acre equestrian park and two golf courses have been proposed along the San Joaquin River within the 100year floodplain. Aggregate mining is a significant use of the San Joaquin Riverbottom. 95 percent of the aggregate used in Fresno and Madera Counties comes from the San Joaquin Riverbottom. At the present rate of consumption, the riverbottom aggregate resources can last approximately 25 years. In the past, aggregate mining was both along the channel and in the channel. This practice of mining the riverbed has now ceased. California's Surface Mining and Reclamation Act of 1975 required that mine operators prepare operational and rehabilitation plans for their sites. Many of the sand and gravel companies operating under old conditional use permits issued by the two counties were not required to preserve riparian vegetation or maintain a riparian buffer. These operations are now in the process of preparing new rehabilitation plans. Finally, road construction is a threat to riparian habitat. The City of Fresno's northern growth has brought about proposals for new freeway and road construction across the San Joaquin River to link residential areas in Madera County to the City.

Preserving and Restoring Riparian Habitat within the San Joaquin River Corridor
The San Joaquin River Committee whose activities I direct was formed in 1985 to support preservation and restoration of San Joaquin Riverbottom plant and wildlife resources. We are a nonprofit organization with over 800 members incorporated under Section 501(c)(4) of the Internal Revenue Code. We perform public advocacy through grassroots organization and public education. We work with local and state representatives to support preservation of riverbottom riparian land and continuation of open space land uses. We helped to qualify Proposition 70 which allocates $5 million for purchase of riparian habitat in the San Joaquin Riverbottom between Freeway 99 and Friant Dam. The San Joaquin River Committee has facilitated discussions with local school districts about utilization of the San Joaquin for outdoor environmental education. We have held numerous public meetings to educate the public on issues affecting the riverbottom. We 81

have covered a wide range of topics from reclamation of gravel ponds for wildlife to San Joaquin River flooding and flood control to public law issues affecting the conservation of open space. An important part of the San Joaquin River Committee's overall program to educate the general public has been special events that bring people to the river. The most successful of these have been canoe floats where we take hundreds of people on day long floats down the San Joaquin. The San Joaquin River Committee publishes a quarterly newsletter. The newsletter has proven effective in educating the public on the importance of preserving riverbottom riparian habitat. While the San Joaquin River Committee has been very effective in "holding the line" on loss of riparian vegetation, the fact that it is primarily an advocacy group has precluded its being able to accept tax-deductible donations of money, land, and easements. We determined that forming a land trust would be beneficial for a longterm program of land preservation. In March of 1988 we formed the San Joaquin River Parkway and Conservation Trust (the Trust), incorporated consistent with Section 501(c)(3) of the Internal Revenue Code. Land trusts across the United States are committed by their charters to long-term management of land and land-based resources and to education about natural resources and the need for their stewardship. Land trusts are locally based, largely self- supporting, and run mainly by volunteers. Land trusts purchase and manage land, negotiate conservation easements, contract with public agencies to manage land, and conduct educational programs. They can act quickly and independently to acquire land when it becomes available, a process that can take months for government agencies. Two wellknown national organizations that perform land trust functions are The Nature Conservancy and the Trust For Public Land. Local community leaders who have agreed to serve as initial directors for the Trust reside in both Fresno and Madera Counties. The directors represent a broad range of interests including developers, environmentalists, lawyers, bankers, teachers, sand and gravel mining, businesses, farmers, cattle ranchers, landowners, and community activists. The Trust has formed advisory committees of local citizens to develop a riverbottom conceptual plan, to develop recommendations on land acquisitions, to research, develop and conduct educational programs, and to fundraise for ongoing activities and specific land projects.

A recent project of the Trust calls for preparation of a conceptual map of a greenbelt-parkway along the San Joaquin River. This conceptual map will map riparian resources, identify sensitive areas for plant and wildlife, and pinpoint critical riverbottom lands for the Trust to acquire. Two methods of protecting riparian lands are acquisition of land fee simple and acquisition of easements. Conservation easements are an attractive way to accommodate a landowner's desire to preserve land while restricting public uses and the public's interest in preserving a greenway. Easements exist at common law as a partial interest in real property which entitles the owner to some limited use of the property or restricts the landowner's use of the property for the benefit of the easement owner. Easements are attractive to land trusts because they preserve land without the higher costs of acquisition. Easements are attractive to landowners because they reduce the appraised value of the land while allowing the land to remain the same. Lands with the least economic value will be the most easily acquired. These lands will include areas prone to floodingthe river corridor itself with its band of riparian habitatand areas mined for sand and gravel that have little potential for other kinds of development. An excellent example of such a property is a 286-acre site under consideration for purchase by the Wildlife Conservation Board. This property includes a small island with low woody vegetation, a series of large irregular-shaped gravel ponds with young riparian vegetation, a large marsh, and reclaimed upland that floods in high water. The efforts of the San Joaquin River Committee and the San Joaquin River Parkway and Conservation Trust have just begun. Hopefully, the lessons learned as we work to preserve and restore riparian habitat along the San Joaquin River will benefit other river corridors in California.

References
Atlantis Scientific. Draft Environmental Impact Report for the proposed Ball Ranch Residential/Recreational Development. June, 1988. County of Fresno. General Plan. County of Madera. General Plan. Counties of Madera and Fresno, and City of Fresno. San Joaquin River Area Reconnaissance Study. June 1986. Coyle, Kevin. Strategies and Tools for Protecting Greenways. November 1987.

82

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

MIDDLE SACRAMENTO RIVER REFUGE: A FEASIBILITY STUDY1


Charles J. Houghten and Frank J. Michny2 The woodlands and other streamside habitat of the Sacramento River's riparian system have been severely reduced within the last century. This riparian habitat and its ability to sustain diverse populations of fish, migratory birds, mammals, and other wildlife have been significantly impacted by water control projects, agricultural developments, and other land uses. The species of particular concern are the western yellow-billed cuckoo (Coccyzus americanus), Swainson 's hawk (Buteo swainsoni), bank swallow (Riparia riparia), wood duck (Aix sponsa), chinook salmon (Oncorhynchus tshawytscha), and the California hibiscus (Hisbiscus californicus), as well as the threatened valley elderberry longhorn beetle (Desmocerus californicus dimorphus) and endangered bald eagle (Haliaeetus leucocephalus). privately owned, and threatened with further loss of habitat with high value to wildlife.

Key Management Issues


Any habitat protection program along the Sacramento River must interface with bank stabilization work conducted by the U.S. Army Corps of Engineers (Corps). In the two lower reaches (south half) of the study area (Map 2), resource issues are being effectively coordinated with the Corps. In the upper reaches (north half) of the study area (Map 3), a comprehensive bank stabilization project placing stone revetment ("riprap") on approximately 35 percent of the river's banks is a more sensitive issue. The principal reasons for concern about this bank stabilization project include: (1) the need has not been clearly demonstrated; (2) salmon spawning may be negatively affected; (3) further losses to riparian vegetation may occur that will reduce the variety of wildlife; and (4) past mitigation efforts generally have been unsuccessful. However, conflicts between bank stabilization and habitat protection can be minimized, particularly if banks were stabilized in an environmentally sensitive manner, and on a "site specific" basis. Other key issues identified during the course of the study included landowner concerns about trespass, and recreationist desires for public access. After a series of meetings with area landowners and representatives of various local, State and Federal agencies, the study concluded that a comprehensive program to protect and restore riparian habitat could be implemented with minimal conflict with other activities.

The Study
In response to directives in House Appropriations Committee Report (No. 99-174) and Conference Committee Report (No. 99-1002) on the fiscal year 1987 Interior and Related Agencies Appropriations Bill, the U.S. Fish and Wildlife Service conducted the "Middle Sacramento River Refuge Feasibility Study." The study identified alternative management actions that could help protect the remaining riparian resources of the Sacramento River between Colusa and Red Bluff, California (Map 1), yet avoid major conflict with other interests or activities. A principal component of the study was the identification of 66 riparian habitat areas within four reaches (or sections) of the 160 kilometer river corridor. The sites, as well as the reaches, were evaluated for their overall habitat value, and prioritized for protection. The report was forwarded to Congress on October 16, 1987, and made available to the public in January 1988.

FindingsOne Possible Approach


The study revealed several feasible approaches for protecting and restoring the study area's riparian resources. One is establishment and maintenance of a riparian zone refuge by State or Federal resource agencies, private conservation groups, or by multiple organizations. The primary goal of land management agencies under a riparian zone refuge concept would be to protect the

Extent of Riparian Habitat


About 6,885 hectares of riparian vegetation remain within the study area. Of this, nearly 1,215 hectares are currently protected by State or Federal agencies. The remaining 5,670 hectares of riparian vegetation, which include approximately 3,645 hectares of woodlands, are

1 Presented

at the California Riparian Systems Conference, September 22-24, 1988; Davis California.

Natural Resources Specialist, U.S. Fish and Wildlife Service, Division of Refuges and Wildlife Resources, Sacramento, California; and Fish and Wildlife Biologist, U.S. Fish and Wildlife Service, Ecological Service Of f ice, Sacramento, California.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

83

existing and currently unprotected 5,670 hectares of riparian habitat within the study area riparian zone. A secondary objective could be the enhancement of areas that are suitable for habitat restoration. A specific refuge boundary has not been proposed in order to provide flexibility in land acquisition and protection techniques. If a refuge is to be established, habitat protection could be accomplished by purchase of fee title or conservation easement, cooperative agreement, or by other means. Areas of intensive recreational use, public access, residential areas, and agricultural areas could be excluded from the program. Lands would be acquired on a willing-seller basis only.

Conclusions
Coordination and cooperation among governmental agencies and area landowners will be a key in the success of a habitat protection program along the Sacramento River. We recommend that the applicability of the riparian zone refuge approach be further investigated for other California riparian systems.

Acknowledgments
The study of the Sacramento River reported here was conducted in coordination with the U.S. Army Corps of Engineers and resource agencies of the State of California. Public input received during the course of this effort was greatly appreciated. We especially thank participants Cathy Osugi, Rich DeHaven, and Barry Garrison, U.S. Fish and Wildlife Service, and editorial coordinator Roberta Burzynski, U.S. Forest Service.

Appropriation
Since the completion of the study, Congress has appropriated $1,000,000 for initial acquisition of lands and establishment of the Sacramento River National Wildlife Refuge. An Environmental Assessment of this land acquisition project is underway and is expected to be completed in Fall 1988.

84

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Map 1Middle Sacramento River Refuge Feasibility Study Area

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

85

Map 2Middle Sacramento River Refuge Feasibility Study Area Refuge Proposal, South Half

86

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Map 3Middle Sacramento River Refuge Feasibility Study Area Refuge Proposal, North Half

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

87

DEVELOPING MANAGEMENT PLANS FOR CALIFORNIA RIPARIAN SYSTEMS 1


Michael Josselyn, Molly Martindale, Dianne Kopec, and Joan Duffield2

Project Scope and Methods


Twelve riparian systems in California between the Sacramento-San Joaquin Delta and the upper Sacramento (near Chico) were studied to develop operation and management plans to restore and enhance riparian habitat (Figure 1). Each of the study areas had recently been acquired by the California Department of Fish and Game (DFG) and were designated as Ecological Reserves or Wildlife Management Areas due to their significance as rare and endangered species habitat. The study involved field inventories of plants and animals on each of the sites, examination of current and historic aerial photographs, and mapping of wetland habitats. In addition, site specific information such as soil maps, parcel maps, and other cultural features was collected. This information was collated into a set of maps and tables for each of the sites. In conjunction with DFG personnel, target species were selected for the area management goals. For each target species, life history and habitat requirements were determined as well as any description of current population status, if known, within the study sites. Based on the habitat requirements, a set of area management objectives was established which would best serve the needs of those species. Each of the sites is affected by human activities. Sherman and Decker Islands have been used as dredge spoil disposal sites. Only a portion of Decker Island is owned by the state and grazing has extended over much of the island. Webb Tract Berms have suffered severe erosion due to boat wakes, dredging activities, and flooding such that of the 285 acres purchased by the State in the 1970's, only 80 acres remain. Woodbridge is surrounded by agricultural lands and must manage its water supply to reduce saturation of surrounding agricultural fields. Riparian areas along both the Feather and the Sacramento Rivers are surrounded by orchards and in some cases the orchards extend into the riparian forest.

Habitat Needs and Management Objectives


The target species for each of the sites is given in Table 1. Most of these species are either state or federallylisted rare or endangered species or are considered species of special concern. Almost all the sites supported six or more of these species, only the Woodbridge site was managed exclusively for the sandhill cranes. Based on the habitat requirements for each of the target species groups, the surrounding land use constraints, and fiscally feasible activities, area management objectives were established for each complex. Some of the key objectives are summarized below. Delta Island Complex

Habitat management must strive for a balance between levee maintenance and provision of mudbanks and floodplains for rare plant species.

Presented at the California Riparian Systems Conference, September 22-24, 1988, Davis, California. Professor of Biology and Director, Romberg Tiburon Centers, San Francisco State University, Tiburon, CA; Research Associate, Romberg Tiburon Centers, San Francisco State University, Tiburon, CA. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

88

Figure 1 Location map for ecological reserves and wildlife areas.


USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

89

Woodbridge Ecological Reserve Expansion of shallow water with low vegetation to support roosting habitat for Sandhill Cranes. Provide extended open water habitat for waterfowl nesting.

Costs
Total management and implementation costs for these project areas range between $250,000 and $600,000 over a 5 year period. These expenditures are necessary if these rare habitats are to serve for the diversity of wildlife now restricted by agricultural and urban development in the Central Valley.

Lower Feather River Complex Restore water levels within the backwater lakes. Restore mixed riparian forest within the floodway. Maintain tall snags and eliminate easements for firewood cutting.

Sacramento River Wildlife Areas Removal of nuisance species such as fig and salt cedar. Planting of valley oak and elderberry. Evaluate downstream impacts of streambank erosion control.

Table 1. Target species for management goals for the Sacramento River, Lower Feather River, Woodbridge and Delta Islands Complexes. COMPLEX Sacramento River PRIMARY SPECIES Valley Oak Forest Valley Elderberry Longhorn Beetle Yellow-billed Cuckoo Swainson's Hawk River Otter California Hibiscus Valley Elderberry Longhorn Beetle Giant Garter Snake Yellow-billed Cuckoo Swainson's Hawk River Otter Sandhill Cranes (Greater and Lesser) California hibiscus California Hibiscus Mason's Lilaeopsis Suisun Marsh Aster Delta Smelt Sacramento Splittail SECONDARY SPECIES California hibiscus Bank Swallow Yellow Warbler Yellow-breasted Chat Purple Martin Long-eared owl Ringtail Turkey Vulture Bank Swallow Tricolored Blackbird Heron Rookery Ringtail

Lower Feather River

Woodbridge

None

Delta Islands

California Black Rail Salt marsh harvest mouse Heron rookery Black-shouldered kite

90

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

SESSION C: RANGELAND AND DESERT RIPARIAN SYSTEMS


The importance of riparian systems in providing for a variety of uses (including wildlife and fisheries habitat, water quality, recreation, and livestock grazing) is well documented in the papers in this and several other conferences and symposia held since 1976. Nowhere are riparian systems more important than in deserts and rangelands. The reason is because deserts and most rangelands occur in areas of low or seasonal precipitation or both, resulting in relatively few riparian areas and greater habitat contrast between the riparian and adjacent upland communities. Recognition of the importance of desert and rangeland riparian systems has resulted in augmented riparian programs on the part of both the Bureau of Land Management, U.S. Department of the Interior, and the Forest Service, U.S. Department of Agriculture. These programs were summarized and commented upon in a report commissioned by Congress (General Accounting Office, Public Rangelands: Some Riparian Areas Restored but Widespread Improvement will be Slow, Report GAO/RCED-88-105, 1988). Controversy, too, has helped drive this new riparian initiative. Probably foremost among the controversial issues is livestock grazing and its effects on riparian areas. Overgrazing of riparian areas was commonplace in the West in the late 19th and early 20th centuries. Even in recent times, when upland rangelands have been improved, riparian zones have continued to suffer from overuse. Traditional grazing systems designed to improve upland rangelands do not usually work in riparian areas. As the papers in this session show, this trend is changing. Management techniques have been designed and implemented to improve the condition of riparian areas. These techniques have ranged from total exclusion of livestock to the implementation of grazing systems designed to improve riparian vegetation. Execution of these actions to improve riparian areas depends not only on land management agencies but on private landowners and ranchers as well. Many ranchers are becoming increasingly aware of the value of riparian areas and are working to improve the management of these areas, as the paper by Flournoy and others demonstrates. Off-road vehicle use and other forms of recreation are impacting desert riparian systems also, especially in California, where urban populations place heavy pressure on nearby areas. Desert riparian systems are afflicted too by the spread of several introduced species of the genus Tamarix at springs and along rivers. The paper by Van Cleve and others discusses management actions taken to ease both of these conflicts.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

91

According to the recent Congressional report (cited above), neither the Bureau of Land Management nor the Forest Service has completed comprehensive inventories of the riparian resources on Federal lands. There is a similar lack of inventory information on the riparian resources on private lands. Clearly, more extensive inventories of riparian areas are necessary. The paper by Gradek and others details the inventory efforts of the Bureau of Land Management in California. An adequate and widely accepted classification system for riparian areas is a necessary precursor to any inventory. Such a system should take into account the important physical and biological components of riparian systems. The azonal nature of riparian systems makes classification more difficult than that of zonal upland ecosystems, but successful approaches have been implemented, as the papers by Bennett and others and Swanson show. Riparian systems tend to be resilient: the presence of water year-round allows riparian vegetation to respond rapidly to management, as the papers by Key and Gish attest. Once improved management has been implemented it is important to allow riparian systems to repair themselves, as the paper by Elmore demonstrates. Too often we have tried to impose our own will upon the systems through the use of structures such as checkdams and streambank revetments, instead of first giving riparian vegetation a chance to do the job. Much work still must be done to improve the degraded riparian areas of the West. This conference hopefully will stimulate land managers, both public and private, to continue working toward improving the condition of these critical ecosystems. John W. Willoughby Bureau of Land Management, U.S. Department of the Interior Sacramento, California

92

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

RANGELAND RIPARIAN SYSTEMS1


Wayne Elmore2

Abstract: The management and recovery of degraded riparian systems is a major conservation issue. Presently there are many grazing management strategies being applied based on the name of the technique with little incorporation of basic stream processes. Managers must understand the exact workings of grazing strategies and the individual processes of each stream before prescribing solutions to degraded riparian systems.

managing riparian systems. In fact, many times wildlife benefits are among the lowest economic value received from riparian restoration. To fully evaluate the benefits and incorporate riparian management into land use plans, I believe that we must go back to basic functions. These functions include: 1. Physical filtering of waterRiparian vegetation can withstand high velocities of water and still remain intact. One of its functions is to slow the flow of water, literally "combing" out sediments and debris. This water purification process also helps to build banks; so channels typically become narrow and deep where once they were wide and shallow. Vegetation, such as grasses, sedges and rushes, lays down under high flows, and literally forms a blanket of protection over the banks. This process reduces bank cutting and aids in deposition of sediments. Where deposition has occurred through time, extensive wet meadows or flood plains develop (Elmore and Beschta 1987). 2. Bank stabilityThe diversity of grasses, forbs, sedges, rushes, shrubs and trees produces a variety of fibrous and tap roots that bind and hold settled soils in place. The binding effect of the roots helps maintain the positive factors of the bank building processes during high flows. A combination of both woody rooted and fibrous species have a reinforcing effect. The woody rooted species provide physical protection to the hydraulic forces of eroding water and allow forbs, grasses and sedges to bind the finer particles. In combination, this diversity of plant species is much more effective in promoting bank stability than is any one species by itself. 3. Water storage and recharge of underground aquifers The aquifers in many areas of the west are going dry and one of the processes of riparian systems is to help recharge a percentage of a given aquifer. For many degraded riparian systems, all flows are contained in the channel and cannot access the banks or floodplains where water can spread. It is widely accepted that we can lower a water table and drain a stored underground aquifer through channelization or erosion. It is not readily accepted, however, that we can reverse that process and store water through recovery of riparian systems and deposition in formerly degraded channels. Riparian systems slow the flow of water and allow it to spread and soak into the banks like a sponge, which raises water

"Riparian" is a word that strikes fear in the hearts of many, anger in some and feelings of peaceful surroundings to others. It is a word that has grown to mean many things to many people, but is rarely understood. It has become an emotional subject that has led to one of the key public land issues in the United States today. Many people are beginning to believe in the old Will Rogers saying "Thank God we don't get all the government we pay for." Early Oregon explorers and residents observed what our riparian areas once looked like. Peter Skene Ogden, after traveling in 1825 through the Crooked River Basin in Eastern Oregon observed willows from side to side across the valley bottom. Most of this scene is now gone. The Indian word "Ochoco," for which our Central Oregon mountains are named, means "streams lined with willows," yet today willows are uncommon. Senior ranchers in Central Oregon tell stories about the problems once encountered gathering cattle in the "thick willow stands" on Big Summit Prairie. The "thick willow stands" have been reduced to scattered clumps. Historic evidence indicates that most riparian zones were then in better condition than they are today.

The Riparian System


In recent years the management or riparian areas has typically been the primary responsibility and interest of wildlife and fisheries biologists (Elmore 1987). Improvements have been primarily judged in relation to habitat for big game, song-birds and fish. But riparian areas are more than just habitat for wildlife. They actually are functioning systems that provide physical filtering of water, bank stability, water storage, and assist in the recharge of underground aquifers along with the adjacent uplands. Wildlife habitat is a product of those functions, and should not be considered as the only emphasis for

1 Presented 2 State

at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

Riparian Specialist, Bureau of Land Management, Prineville, Oregon.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

93

tables. When banks rebuild through filtering of sediments, they increase the area for water absorption and improve recharge of aquifers by allowing gravity to work on the stored waters. Upland areas must not be excluded in this paper because they are an integral part of the riparian system. Overland and subsurface flows also influence sediment loads, water cycles, and recharge of aquifers. Other processes I have observed in Eastern Oregon riparian systems, that have shown a substantial ecological improvement, include increases in the base flow (minimum flow level, i.e. the discharge to which the stream returns after storms or snowmelt periods), reduction in buildup of ice, and physical filtering of sediments by ice. Almost all of the processes I have observed as negative in our stream systems today become positive factors when those streams are in good ecological condition. The education we transfer to the managers and users of our natural resources must contain this basic information.

percent and again at 85-90 percent utilization. The third year we rest the pasture and, hopefully, no use occurs. In analysis we can see that we are basically losing three years of growth on willows and only getting two years of growth back. However, at the same time we are meeting the physiological need of the sedges, rushes and grasses. There are many things we could do to solve this problem. One is to restrict riparian utilization during the seed ripe treatment to 50 percent or less. Another is to make the riparian area a separate pasture. A third is to add more pastures to achieve more rest, or finally we could exclude the stream from grazing. The point is you must know what your proposed management is going to do, and how it will work in each individual stream system. Other grazing systems that we commonly use in Eastern Oregon are deferred (graze after seed ripe every year) and early or spring grazing. Deferred or seed ripe use every year can quickly remove small shrubs from streams systems because of heavy riparian utilization, but can also increase sedge and rush communities in wide low gradient valley systems. Early or spring use every year can be beneficial to riparian system recovery, but many times this system can be detrimental to upland grasses if grazing always occurs during the critical part of the growing season (when flower stalks emerge from basal bud). It is very apparent that utilization of riparian vegetation should not be a major concern unless it affects stream function. This occurs commonly with deferred grazing systems on sites where regrowth is limited and in the use of three pasture rest rotation where shrubs are needed for bank stability and sediment filtering. Figure 1 is a simplified look at how we try to analyze our riparian systems and proposed management techniques. Every management strategy exerts an amount of stress on our riparian systems. The ability that each stream has to handle this stress depends on its own natural stress or sensitivity. Some streams with high natural sensitivity such as those with bentonite soils and high erosion potential are immediately in the caution area and probably can stand little, if any, management stress (human influences). Others that are low gradient with sandy loam soils, for example, can recover under much higher management stress. In our evaluation, the stress of management must not be confused with livestock numbers. Often, for streams in poor condition, livestock reduction was proposed as a solution. However, no recovery in the stream occurred. It was not the numbers of livestock that was the problem, but the management strategy. For example, Bear Creek in the Prineville BLM District previously had 73 animal use months (forage needed to sustain a cow for one month) of grazing under a season long strategy. This was more management stress than over 300 animal use months

Management Evaluations
Understanding riparian system functions is essential to their management. In management applications as we endeavor to restore streambank conditions, we are often applying techniques based primarily on the name given the technique and not on what that technique actually does. For example, the three Pasture Rest Rotation grazing system works very well in Central Oregon on low gradient streams that are primarily grasssedge-rush sites, but can be a disaster on streams that need shrubs for bank stability. If we look "inside" this grazing system, we find that it was designed to fit the physiological needs of grass plants and not riparian shrubs. If we look even closer at what happens under this grazing system in desert rangelands, we can see why shrubs generally decline. The first year we graze the pasture early during the growing season. The second year we graze the pasture after upland grass seeds ripen (usually mid-July), and the third year we rest the area from grazing. During the spring use period, we receive little if any utilization on willows by livestock. Upland grasses are green and growing, providing a much more palatable forage source than shrubs. During the second year, the common utilization rate for upland grasses in this grazing system is 60 percent. These grasses are now dry and unpalatable and by the time we have achieved the desired 60 percent utilization on uplands, we have gotten 80 to 90 percent utilization on riparian zones. Our observations in Oregon show livestock will begin using the current annual growth on willows during the seed ripe treatment (mid-July through September), when riparian utilization reaches 45 percent. They will increase their use on shrubs again at approximately 65

94

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

now exerted during early spring grazing. As a result of decreased stress, the creek is making significant improvement with a four fold increase in grazing use. There are many other examples in Oregon and throughout the Great Basin, as exhibited in research work by Bill Platts and others (Platts and Raleigh 1984; Platts and Nelson 1985). Exclusion of livestock is a management strategy that has been proven to work in inducing the recovery of riparian areas. It continues to receive a lot of criticism from many managers and users of the public lands for several reasons. Some of them are: expense of fence construction, maintenance, wildlife concerns and livestock water. However, if we look at many streams in poor ecological condition they have become, in effect, upland exclosures. The attractant nature of streams to livestock during summer grazing periods many times excludes livestock use on 90 to 95 percent of the adjacent upland areas. What we typically observe with streams in poor ecological condition are all of the negative things we receive with improper grazing concentrated in one area. At the same time, we receive none of the positive factors of grazing in the upland areas where they were planned. We are also, I believe, many times comparing exclusion of livestock to improper grazing and not comparing it to proper grazing. There should be three scenarios in our evaluations, not two.

Conclusions
We must begin to realize that we can look at things in a different way and that changes in management can provide recovery in our stream systems. The benefits from those changes far outweigh the costs of continuing with our present practices. The watershed, not just the stream system, must be our focal point. As our energy and dollars focus on restoring degraded streams, we also have to look at the uplands. We cannot forget that the speed and clarity that water comes off our uplands has a big impact on what happens in the stream system. If our goal is a higher quality and quantity of useable water, then the other 98 percent of our rangelands must be a part of our program. We are at a critical time in the management of riparian areas and associated uplands. "Members of the livestock industry can provide leadership in understanding and solving complex riparian questions (Elmore and Beschta 1987). We must begin to look at both private and public lands because riparian areas have never been able to tell the difference in ownerships, only in management. If we don't change our management, we will either lose the benefits of our natural resources, or we will lose the flexibility to manage for multiple use. The American public is concerned about useable water quality and quantity as evidenced by the recent Congressional override vote on the President's veto of the Clean Water Act. The public will demand more from the management of our natural resources and we must start now to meet those demands. Just remember, you will never see a picture of a degraded riparian zone on a calendar so why should we have them in our landscapes. Riparian management full stream ahead.

References
Elmore, W., 1987. Riparian Management: Back to basics. The Public Lands During the Remainder of the 20th Century. Univ. of Colorado School of Law. June 8-10, 1987. Elmore, W.; Beschta, R.S. 1987. Riparian areas: perceptions in management. Rangelands Vol. 9 (6) 260-265. Platts, W.S. and Raleigh, R.F., 1984. Impacts of grazing on wetlands and riparian habitat, In: Developing Strategies for Rangeland Management. National Research Council/National Academy of Sciences. Boulder, CO: Westview Press, 1105-1117. Platts, W.S. and Nelson, R.L. 1985. Impacts of rest-rotation grazing on streambanks in forested watersheds in Idaho. N. Amer. Jour. of Fisheries Management, 5: 547-556.

Figure 1Natural stress or sensitivity of streams vs. management stress. Factors like soils, gradient, water column, climate, etc. must be considered when designing management strategies for system recovery.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

95

USING STREAM CLASSIFICATION TO PRIORITIZE RIPARIAN REHABILITATION AFTER EXTREME EVENTS1


Sherman Swanson2 Abstract: Historic use of many stream riparian areas and associated watersheds has impaired the capacity of riparian vegetation and floodplains to reduce stream energy and trap sediments. As low-gradient streams with erodible banks increase in width and change their pattern, they approach a threshold of instability. Once a stream exceeds a threshold, it must proceed through a process of geomorphic gully evolution that includes degradation, widening, and aggradation phases. Opportunities for enhancing and maintaining favorable conditions of stream morphology and associated riparian values vary throughout this process. The highest priority stream reaches for watershed, riparian, and stream management are those approaching the threshold. After the degradation phase, the marginal reaction to management input increases as the gully widens. Riparian grazing can be managed in a variety of ways to avoid detrimental effects. A useful alternative to a riparian exclosure is a riparian pasture that can be managed for optimum riparian resource values. However, when approaching a threshold (Van Havern and Jackson 1986), there can be substantial effect during a flood if either or both of two conditions occur: 1. The cohesiveness of stream channel materials weaken significantly; or 2. The forces impinging on the streamchannel materials increase because of some change in the cross-valley profile that confines the flood wave. Historical land management has often created both of these conditions. Furthermore, inappropriate management of mining, road building, timbering, fire, or grazing has caused many watersheds to release water and sediment at substantially increased rates. Increased flows force the stream to adjust and they may exceed the capacity of the natural or stressed stream channel to convey them without significant alteration. Although streams approach and exceed thresholds of instability under natural conditions, it normally requires dramatic geologic or climatic change for a large number of streams to approach threshold within a time period as short as man's influence on the West. It seems inappropriate therefore, to attribute the inordinately devastating effects of rare but natural events to "acts of God". This paper uses concepts developed from stream classification (Rosgen 1985) to describe the role of riparian vegetation and floodplains in maintaining stream channel morphology in low-gradient streams. From these concepts is drawn an approach for prioritization and management of such streams. Although many of the principles apply broadly, the management field of livestock grazing is emphasized.

The pattern of settlement and the history of use of the American West has left today's natural resource managers with many riparian problem areas and horror stories. Many stream environment zones that have degraded are now unraveling and have been doing so for decades Cottom and Stewart (1940). Some have not yet reached the threshold leading to collapse. Others are progressing through the process of geomorphic and ecological recovery. As years pass, additional stream reaches succumb to the convergence of a major runoff event and an approach to a threshold. This happened to many streams in the early 1980's when successive winters produced abnormally high runoff that each year prolonged the period of high flow. Active stream-channel dimensions conform to the bank-full flow that typically represents the normal high water mark (Wolman and Miller 1960). This bank-full flow comes only once or a few times in most years. It is effective in forming the channel that conveys it because it represents the greatest cumulative energy level. Larger flood events last for too short a time to generate much effect even though their energy level is extreme for a short time. Low-flow events lack the energy even though their duration is substantial.

Function of Riparian Vegetation in Stream Morphology


At the Sheldon Antelope Range in Northwestern Nevada, Nebraska sedge (Carex nebraskensis) dominated communities have an average of more than 2 meters of roots and rhizomes per cubic centimeter in the top 10 centimeters of the soil profile (Manning 1988). It is no wonder that it and other broad-leaved sedges have gained a reputation for stabilizing sediment and binding stream-bank soil (Youngblood and others 1986). Although other species of herbaceous plants may not have as great a root-length density, it is not uncommon to

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Extension Range Specialist and Research Scientist, University of Nevada, Department of Range Wildlife and Forestry, Reno, Nevada. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

96

see stream banks that are stable because of the tough sod produced by plants that thrive in the moist conditions found with a high water table. Willows and other woody riparian species have also achieved a measure of notoriety for their role in stream-bank stability. Besides providing cohesiveness to otherwise erodible alluvial materials, vegetation provides roughness that increases friction at the water-land interface. This decreases velocity and decreases the energy available for doing work including detachment of channel materials and transport of bedload or suspended sediment. The filtering effect of riparian vegetation is partly responsible for deposits of fine fertile soils on many floodplains such as mountain meadows. Within the active channel it is also instrumental in the process of narrowing streams that are recovering from bank erosion. It is natural for streams on low (<1.5 percent) gradients with floodplains to meander (especially C6, C4, and C3 stream types (Rosgen 1988)). This involves a balance of erosion on the outside turns and deposition on the inside turns. In order for streams to remain stable, the rate of these two processes must remain in approximate equilibrium. If the outside erodes faster than the inside captures and stabilizes sediment, a narrow deep stream that could provide tremendous habitat for cold-water fish may become wide and shallow. As the stream widens, the stream pattern changes accordingly. Streams tend to form meanders that are approximately 7 to 10 times as long as the stream is wide (Leopold and others 1964). Characteristically, as a stream widens it breaks through meanders and the broad sweeping curves of the new channel lead to decreased stream length. Sinuosity is inversely related to channel gradient for a given reach of stream maintaining constant elevation at the top and bottom ends. Therefore, as the stream straightens, the gradient and velocity increase. The total energy is thus expended over a shorter length of channel. It can exceed critical shear and accelerate erosion. Figure 1The broad floodplain of a narrow, deep sinuous channel dissipates flood energy allowing vegetation to build and stabilize stream banks. Tractive force is directly related to depth of flow and slope. Therefore as a stream floods it has increased energy available to do work (erosion) on the stream channel largely in proportion to the increase in depth. A stream that can spread out over a broad floodplain increases depth only a small amount during a flood event therefore it can withstand floods of tremendous magnitude with little erosion. Such streams will generally deposit fine sediment on the floodplain and build stream banks during flood events. As stream reaches with broad valleys capture sediment, they gradually steepen. Under natural conditions, the stream valley may become too steep for meander maintenance Patton and Schumm (1981). When meanders begin to cut and the stream straightens, the concentrated energy can downcut the channel by exporting channel materials. This can initiate a nick point that develops into a headcut (fig. 2) and proceeds upstream, assuming a life of its own. Any net export of channel material causes the stream to lose some accessibility to its floodplain. As the floodplain loses its ability to dissipate flood energy, the energy of the confined and therefore deeper stream energy accelerates the process of downcutting until the stream reaches a gradient that is low enough, or the new channel materials are coarse enough, to stop downcutting. At this point the stream approaches local base level. A totally confined stream (gully or arroyo) on a low gradient (<1 percent) is labeled F by Rosgen (1988). Initially the stream width is the same as the gully-bottom width (fig. 3), the old floodplain is a terrace, and there is essentially no floodplain. Therefore energy is very concentrated and high water continues to do work by eroding the gully walls.

Function of Floodplains in Stream Morphology and Gully Evolution


One of the characteristics of a narrow deep sinuous (C6) (Rosgen 1985) stream (fig. 1) is that the surface of the water is near the surface of a broad flat floodplain. The high water table provides abundant water to the vegetation that in turn provides the bank stability upon which stream morphology depends. The broad flat floodplain is necessary for dissipation of energy during flood events.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

97

the gully. It then would be labeled a C type by Rosgen (1985). The farther apart the gully walls become, the more the floodplain can dissipate energy and the more effect streambank vegetation can have in controlling the morphology of the active channel (fig. 4). As gully banks recede, there will eventually be aggradation on the expanding floodplain. Then floodplain widening can proceed under the dual influence of gully bank erosion and filling of the trapezoid-shaped gully. The gully banks define terraces that eventually may again become floodplain if the gully fills sufficiently. At any point in the recovery, the aggrading sediments may again be cut by a new cycle of gully evolution. This cycle of aggradation and degradation has occurred repeatedly in some mountain meadows since the Pleistocene (Wood 1975). The time between cycles depends on a combination of factors including sediment supply from the headwaters, size and shape of the valley, climate, etc. Modern man has triggered the degradation phase of this cycle prematurely in thousands of locations by land use practices. Roads and trails on floodplains are notorious for their effects on streams because of their tendency to help the stream cut through meanders. Some roads and trails have been captured by floodwaters to become stream channels. Their straight path allowed the tractive forces of floods to excavate a completely new channel, a gully. "Improved" roads may accomplish the same effect by covering part of the old floodplain area with road-fill material. This not only removes potential valley bottom for the stream to meander across but also confines floodwater and thereby increases its depth and energy.

Figure 2Headcuts concentrate the energy of flowing water, thereby accelerating erosion, downcutting, and confinement.

Figure 3Initially downcut streams are as wide as the gully bottom and have no floodplain. The water table that previously supported dense vegetation on the old floodplain is lowered as a result of downcutting. Riparian vegetation is then replaced by more xeric species such as sage brush (Artemisia tridentata) and cheatgrass (Bromus tectorum). The over-steepened gully walls typically remain unvegetated or lightly vegetated because of their natural instability and xeric soil conditions. Even as vegetation colonizes the water edge at the bottom of the gully wall, it is subject to extreme tractive force during high water because of the confinement of the stream. Therefore, the active channel in the bottom of a gully soon achieves a high width/depth ratio (10- 40). It stays wide and shallow until the gully walls erode apart far enough for there to develop a useful floodplain in the bottom of

Figure 4 The emerging floodplain of a widening gully dissipates energy and promotes vegetative stabilization of the active channel.

98

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Other land-use activities may also produce gullies. Many stream valleys are used for transporting logs. In previous decades, the stream itself was sometimes the vehicle. The grazing effects of concentrated livestock in riparian areas is widespread where grazing management has not prevented distribution problems. Livestock grazing (or abrasion by logs) on stream banks can have the effect of caving in the overhanging banks (fish cover) that otherwise form on low-gradient meandering streams with erodible soil. As the stream banks erode from the physical effect of trampling or because of weakened root systems, the opposite bank must be able to capture and stabilize sediment in order to maintain the equilibrium and the narrow channel. If residual vegetation is not available during the period of high water, or if the grazing and trampling effects are too great, the net effect is first widening and then, if the gradient of the valley becomes too steep, downcutting. Many if not most streams located in wide valley bottoms have downcut to some extent in the last century and a half. The tremendous amount of sediment coming from these eroded stream banks and gully banks has in turn caused additional problems downstream.

cycle will yield far greater benefits over time than would comparable input invested during another phase. Highest Priority Stream Reaches The highest priority streams are the ones that still have and use their floodplain, especially if the use of it could be lost through downcutting (figs. 1 and 2). Streams that still rely on stream bank vegetation growing at the same or nearly the same level as the floodplain will be most likely to respond to appropriate riparian grazing strategies. This is in part because of the availability of water and the vegetative resilience that comes with water availability. It is also due to the energy dissipation influence of the floodplain. If the stream bank is composed of fine-grained erodible soil, especially sand, silt, loam, or fine gravel, and if the stream is or was highly sinuous (C6, C4, and C3), it is probably most dependent on bank vegetation. If the stream has begun to downcut, it may be approaching a threshold of instability which, once exceeded, may require a long period of gully downcutting, widening, and filling to duplicate present riparian values. Proper management is especially critical in stream valleys that are long and deeply filled with erodible alluvium that has consistently depended upon streambank vegetation for streambank and meander integrity. Once headcuts form, they are very difficult to heal vegetatively. The time to act is before the threshold is exceeded and the nickpoint initiated. Lowest Priority Stream Reaches The lowest priority streams are the ones that are unlikely to respond to management even if they are the ugliest and even if they were once the prettiest (fig. 3) Where a stream has downcut and is totally confined in the bottom of a gully, stream energy is concentrated and management inputs are likely to be wasted. It is common for land managers to remember or presume how the meadow or streamside floodplain used to look and to want to refill the gully. High check dams are a commonly used method for attempting to achieve this. Predictably these normally wash out. As it approaches local base level, a gully progresses through its natural evolution of widening. Behind dams, widening is accelerated because energy is redistributed against the bank at an elevated stage. Designers who recognize this often prolong the life of a dam by extending the keyways well into the banks. The concentrated energy dissipation at the dam is also a hazard if the dam is too steep on the downstream side, if the downstream banks are not adequately protected, or if the plunge pool is inadequately armored. If the

Prioritizing Land Management Settings in Evolving Landscapes


Land managers must accept the history of land use that has preceded them. By understanding that history and the physical and ecological attributes of it, they can better appreciate the trend of their landscapes and the potential of those landscapes to respond to management. Effective land managers recognize the limits of their financial, physical, temporal, and managerial resources. They focus attention on land management practices that will most significantly improve resource values over some future period. In an evolving landscape, it is not useful to compare what could be with what is. One must instead compare what could be assuming option A, with what might be assuming option B. This must be done in individual settings to determine if the possible or proposed actions will be worthwhile. It also must be done in many settings simultaneously to determine where and how limited resources can do the most good. Economists term this the best marginal reaction. Major problems or opportunities are commonly concentrated in small areas of a land unit, along certain roads, stream reaches, etc. Here is the place to begin prioritizing. However, care must be taken to avoid the approach of simply attacking that which is most ugly. Considering the evolution of stream valley degradation and aggradation discussed above, it is clear that land management input invested during certain phases of the

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

99

dam is effective at redistributing flood waters over the old floodplain, some of the water must at some point reenter the gully. The concentrated energy dissipation at that point commonly initiates a headcut that can also bypass the dam. The hazard of this may increase as flood waters attain higher elevation behind a dam that is filling with vegetation-stabilized sediment. It is possible to capture significant resource values, at least in the short run, with check dams in gullies (Swanson and others 1987). However, the financial cost can be high and the risk of failure increases with the quantity of water available to do work. The best application of check dam treatment is high in the watershed on small gullies that have reached local base level or where bedrock protects the lowest of a series of dams from an upstream migrating headcut. In general, low structures (1/10 to 1/4 of the active channel bankfull height) are preferred to high structures (1/4 to 2/3 gully bank height). For a discussion of how to choose the correct design for fish habitat improvement structures for particular stream types, refer to the work of Rosgen and Fittante (1986). They point out that many stream "improvement" structures, when placed in inappropriate stream types, cause more damage than benefit. Any of a variety of structures can produce benefit if properly used in the correct stream type. Another common response to gully erosion, when it results from livestock grazing, is dramatically altered livestock management. Although protection of the riparian vegetation colonizing the gully bottom may provide some decrease in the width/depth ratio of the active channel in the bottom of the gully, and may slow the rate of gully widening, the effects are minimal. The opportunity for benefits to exceed costs are lowest in the early phases of the degradation/aggradation cycle discussed above. The marginal reaction of an investment in intensive livestock grazing management increases as the gully bottom widens.

widening (erosion) becomes more important. Sometimes the benefits of even a little riparian management and riparian vegetation along the bottom of a narrow gully prove worthwhile. However, if sediment is a big problem, the marginal reaction of investments to prevent the gully in the first place could have paid for some rather intense management. Also, such receiving streams will likely have suffered significant alteration from the sediment received after initial gully formation. Some stream types (such as flat gradient (<1 percent) gravel or sand bed streams with fine-soil banks, C3 and C4 (Rosgen 1985)) substantially increase bank erosion after an input of sediment. Sediments deposited in bars occupy channel capacity and force the stream to redistribute energy against its erodible banks. Sediment also fills reservoirs and may become trapped in coarser gravels that must be clean to provide adequate fish spawning habitat. Other receiving streams can tolerate substantial input of sediment without significant alteration of channel morphology or resource values. The sediment is simply routed downstream to larger streams or rivers.

Increasing Priority Stream Reaches A dramatic shift in the potential of a gully bottom stream to produce a narrow stream channel conducive to cold-water fish appears to occur at about the time the gully bottom becomes wider than the active channel (fig. 4). At this time the floodplain inside the gully has begun forming and can begin to dissipate some flood energy. Riparian management and riparian vegetation then become significantly more important. The marginal reaction of investments increases most with gully widening if the benefits are measured on site. These benefits include improved fish habitat, riparian vegetation, and aesthetics (fig. 5). To the degree that sediment is a concern downstream, the rate of gully

Figure 5 Gullies that are old, wide, and well managed become valued again for riparian vegetation, fish, and wildlife habitat, water and sediment storage, and aesthetics.

Grazing Management for Riparian Benefits


Livestock distribution is the number one grazing problem in the western United States. The heart of the problem is commonly over used riparian habitat. Controlling utilization is a central precept of grazing management. However, it must be recognized that this is not simply controlling the number of grazing animals or the length of time that they are in a pasture.

100

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Typically cattle graze certain species and certain areas before they graze others. The species and areas are likely to change from season to season and the effect of grazing and trampling has a different effect on different species and areas at different times. This allows a careful observer to identify problem areas and practices that cause unacceptable damage. The manager can then use a variety of livestock management tools to avoid the problem. Grazing systems specify the season, the length of time, and the number of animals that can graze a pasture. Often grazing systems specify a rotation pattern so that periods of grazing that are in some way detrimental do not occur every year. Range improvements, such as water development and vegetation manipulation, that encourage livestock to increase use of previously under-utilized areas can also take pressure off riparian areas. Perhaps the most direct means of control is a well maintained fence. Fences, however can serve diverse purposes. The design of a fence means a great deal to both the use of the area and the cost of the fence. Use of riparian exclosures has made it obvious that stopping bad grazing practices can produce tremendous benefits to streams and to fish and wildlife habitat (Platts and Rinne 1985). From riparian grazing research (Platts 1986) and accumulating experience (Elmore and Beschta 1987), it is also becoming apparent that improved grazing practices can produce improved riparian and stream conditions. Improved grazing management can do this without placing an exclosure fence in a recreation or wildlife use area. A useful practice especially along streams with broad floodplains and expansive areas of abundant riparian vegetation is the riparian pasture (Platts and Nelson 1985). This avoids the problem of cattle concentrating in a small riparian part of very large pastures and allows grazing managers to efficiently tailor riparian grazing to optimize riparian values. Some riparian grazing management techniques such as grazing systems and seasons of use that are appropriate for particular settings are discussed by Elmore in this volume.

References
Cottom, W. P.; Stewart, G. Stewart. 1940. Plant succession as a result of grazing and of meadow desiccation by erosion since settlement in 1862. J. For. 613-626. Elmore, W.; 1989. proceedings. Rangeland riparian systems. These

Elmore, W.; Beschta, R. L.. 1987. Riparian areas: Perceptions in management. Rangelands 9(6):260-265. Manning, M. E. 1988. The ecology of rooting characteristics of four intermountain meadow community types. MS Thesis, Univ. Nevada Reno, Reno, Nevada. 92 pp. Patton P. C. ; Schumm S. A. 1981. Ephemeral-stream processes: Implications for studies of Quaternary Valley fills. Quaternary Research, 15:24-43. Platts, W. S. 1986. Riparian stream management. Transactions Western Section The Wildlife Society 22:90-93. Platts, W. S.; Nelson R. L. 1985. Will the riparian pasture build good streams? Rangelands, 7(1):7-10. Platts, W. S.; Rinne, J. N. 1985. Riparian and stream enhancement management and research in the Rocky Mountains. N. Am. J. of Fisheries Management, 5:115125. Rosgen, D. L. 1985. A stream classification system. Riparian ecosystems and their management: Reconciling conflicting uses - First N. Am. Riparian Conference (proceedings). Tucson, Ariz. April 16-18. USDA For. Serv. Gen. Tech. Rep. RM 120, p 91-95. Rosgen, D. L.; Fittante, B. L. 1986. Fish Habitat Structures - A selection Guide using stream classification. Fifth Trout Stream Habitat Workshop. Lock Haven University, Lock Haven, Penn. Aug 11-13. Swanson, S.; Manning, M.; Franzen D. 1987. Rodero Creek: Rising water on the high desert. J. Soil and Water Conservation, 42(6):405-407. Wolman, M. G.: Miller, J. P. 1960. Magnitude and frequency in geomorphic processes. J. Geology, 68:54-74. Wood, S. H. 1975. Holocene stratigraphy and chronology of mountain meadows, Sierra Nevada, California. PhD dis. California Institute of Technology, Pasadena, California. Van Havern, B. P.; Jackson, W. L. 1986. Concepts in stream riparian rehabilitation. Trans. 51st. N. Am. Wildl. and Nat. Res. Conf.:280-289. Youngblood, A. P.; Padgett, W. G.: Winward, A. H. 1985. Riparian community type classification of eastern Idaho - western Wyoming. USDA Forest Service Intermountain Region R4-Ecol-85-01. 78 p.

Acknowledgements
I thank Dave Rosgen for his help on stream classification and dynamics.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

101

TEN YEARS OF CHANGE IN SIERRAN STRINGER MEADOWS: AN EVALUATION OF RANGE CONDITION MODELS1
Barbara H. Allen2 Abstract: Grazed Sierra Nevada stringer meadow systems were sampled on Blodgett Forest Research Station in northern California between 1977 and 1987 to determine cattle use, and to examine changes in production and species composition over time. Utilization of meadow species averaged 61 percent over the 10 years, but use increased to more than 80 percent utilization after 1985. Production averaged 2733 kg/ha but has significantly declined in recent years. Relative species composition has not changed, nor has total vegetative cover between 1979 and 1986. Range condition models based on changes in species composition were not useful for assessing these stringer meadow systems. Managers should instead base livestock management on stream bank conditions and meadow productivity. Montane meadows in California range in size from a few square meters to several hundred hectares interspersed through every forest type (Allen 1987). Meadows provide forage and cover for an estimated 260 animal species, including at least 12 sensitive, rare, or endangered animals (Timossi 1988). Water and recreation from meadows are valuable resources (State of California 1988). Although small in extent, meadows in the Sierra provide up to half the summer forage for livestock. Historically, heavy stocking rates and consequent overuse changed meadow hydrology, species composition, and production. Although Federal laws have regulated livestock use since 1891, many meadow systems are still in poor condition (Ratliff 1985). Range condition can be defined as "the current productivity of a range relative to what that range is naturally capable of producing" (Society for Range Management 1974). Most range condition models use species composition as the foundation for assessing condition or "range health". For example, the quantitative climax approach (Dyksterhuis 1949), compares current species composition to climax species composition. The Forest Service's score card approach adds vegetation production, soil erosion, and soil cover criteria to species composition when assessing condition (USDA 1969). Generally, productivity and forage values are assumed to be highest when the rangeland plant community is near climax, and lowest when the vegetation is in an early seral stage (Sampson 1952). Early studies in the Midwest showed that grazing can cause changes in species composition through selective use of specific plant species (Dyksterhuis 1949, Ellison and others 1951). Continuous grazing at heavy stocking rates, or during the wrong season, decreases palatable (and also climax) plant species and increases invader and increaser species (Dyksterhuis 1949). Changes in rangeland vegetation resulting from overgrazing allow condition to be expressed in condition classes, generally excellent, good, fair, or poor' (Sampson 1919). Reduction of grazing can result in a return of climax species and improved range condition (Branson 1985). Improvement can also be quantified by condition class the changes retracing the path of range deterioration (Heady 1975). These concepts were applied to rangeland ecosystems throughout the western United States. This paper reports a study to evaluate long-term changes in production and species composition on grazed stringer meadow systems, and to determine the value of range condition models for meadow management.

Methods
This study began in 1976 as part of a larger mixed conifer forest grazing study (Kosco 1980) at Blodgett Forest Research Station, near Georgetown, CA, at 1500 meters elevation on the west slope of the Sierra Nevada. The study site's Mediterranean climate has cool, wet winters and little summer rainfall. Precipitation is concentrated between October and March, averaging 171 cm/yr. The range between 1977 and 1987 included a low of 58 cm and a high of 279 cm. Winter high temperatures averaged 8C, while daytime summer temperatures averaged 28C. Meadows comprise less than 1 percent of the study site. Some herbaceous vegetation grows in disturbed areas such as clearcuts, skid trails, and along roads, but most grows in stringer meadow systems, which follow natural drainages (fig. 1). Some areas remain wet throughout the year, while slightly higher areas dry out as the summer progresses. This results in a visual (as well as statistical) difference in botanical composition within and between meadows.

1 2

Presented at the California Riparian Systems Conference, September 22-24, 1988; Davis, California. Assistant Professor of Forestry and Resource Management, University of California, Berkeley, California. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

102

Figure 1Stringer meadows vary from a few meters across to about 40 m, following the water courses in the mixed conifer forest.

Cattle grazed the study area between May 30 and September 30 every year. The study area is not fenced; the cows and calves are free-roaming, following water courses and forage supplies in meadows and clearcuts. The permittees distribute animals over their allotment, which includes the study area, by riding, salting, and culling of animals. Thirty-five cow/calf pairs (140 animal unit months) were allocated to graze the study area.

Herbaceous production was determined from clipped 1/16 m2 quadrats taken from the center of each caged plot. Clipping occurred at peak above-ground production, generally mid-September. Each spring the cage was moved to a new location near the previous one to reduce cage effects on production. Utilization of herbaceous vegetation was estimated by clipping a 1/16 m2 grazed plot paired with each caged plot used for the production estimates. Utilization was calculated from the equation: Utilization = (UGHW GHW/UGHW) x 100

Production/Utilization Fourteen square meter, paired, caged and uncaged plots were used to determine meadow production and utilization between 1977 and 1987. Visibly distinct types based on species composition were recognized during a reconnaissance survey of the study site. Cages were placed to characterize meadow vegetation on the whole study area. This resulted in 14 cages being placed on 8 meadow systems.

Comparisons of meadow production and utilization were made using analysis of variance (Zar 1984) over all years. Multiple comparisons between means for years were examined using Duncan's multiple range test at alpha of 0.05 (Norusis 1986).

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

103

Species Composition Species composition was recorded using a 10-point frame within each cage before it was clipped in 1979 and 1986. Fifty points per cage were recorded; totaling 700 points per year on the meadows. Species composition between 1979 and 1986 was compared using ANOVA and the nonparametric MannWhitney test for differences between means (Norusis 1986). Permanent Exclosure A 3.5 by 5 m exclosure was erected in 1977 to provide information on long-term changes in species composition without grazing. The exclosure represented Forest-wide meadow species composition with a mixture of grass, forbs, rushes, and sedges. The exclosure was sampled with 300 points using a 10-point frame in 1979 and again in 1986. Significant differences in plant groups were compared between years using the binomial distribution (Zar 1984). Exclosure production was sampled from 5 clipped plots and compared to average meadow production on caged plots in 1988.

Species Composition Meadows on Blodgett are relatively depauperate in number of species. Only five species of grasses, several species of sedges and rushes, and 11 forbs were recorded (Table 1). Total vegetative cover averaged 88 percent in 1979 and 84.4 percent in 1986. Grasses do not dominate these meadows. Grass composition totalled 20.8 percent in 1979 and 13.2 percent in 1986. Forbs form the larger category with Mimulus primuloides dominating in 1979 and 1986 at 27.3 and 21.5 percent relative cover, respectively. Except for an increase in sedge cover, no category differed significantly between 1979 and 1986. No statistically significant difference occurred in percent bare ground/litter. Following Ratliff (1985), species were designated as decreasers, increasers, or invaders (Table 1). Decreasers declined from 19 to 14 percent, while increasers also declined from 60 to 48 percent from 1979 to 1986. These statistically insignificant declines were largely offset by the increase in sedge species which may act as increasers or decreasers in meadows, depending on species. Finally, the meadow species were assigned to categories of desirable, less desirable, and undesirable based on their relative palatability to cattle using standard USDA Forest Service scorecards for meadow species (USDA 1969). None of these categories changed significantly between 1977 and 1986. Table 1 - Relative foliar cover of Blodgett Forest meadow species in 1979 and 1986.1
Percent Relative Cover2 Species Agrostis thurberiana Agrostis sp. Danthonia californica Muhlenbergia filiformis Panicum lanuginosum Carex spp. Juncus oxymeris Juncos & Luzula spp. Trifolium repens Mimulus guttatus M. primuloides Hypericum anagalloides Viola sp. Unknown forbs Moss species 1979 3.1 5.2 8.9 2.8 0.8 6.6 19.3 2.8 7.4 1.8 27.3 10.8 1.9 0.3 0.6 1986 4.3 5.9 1.0 2.1 23.6 13.8 5.3 1.9 21.5 12.7 2.6 1.6 0.3 D,Inc,Inv3 Inc Inc Inc Inc --D/Inc D D/Inc/Inv Inv --Inc Inc Inc -----

Results
Changes in Production and Utilization Average above ground biomass production on Blodgett Forest meadows varied significantly among years between 1977 and 1987 (fig. 2). In general, years prior to 1985 produced significantly more standing crop than after those years, except for 1977 which followed 3 consecutive years of drought. Mulch, the herbage remaining at the end of the grazing season, significantly changed during the study (fig. 2). With a pattern similar to production, mulch averaged 1046 kg/ha prior to 1984, but only 581 kg/ha between 1984 and 1987. The three wetter meadow types on Blodgett produced 4210 kg/ha in 1987. Dry meadow types produced an average of 1590 kg/ha, while nine moist types produced an average of 2495 kg/ha. Utilization, or the proportion of vegetation consumed by grazing animals, differed significantly among years during the study (fig. 3). Utilization averaged 55 percent prior to 1984, and averaged 71 percent after that, peaking in 1986 at 82 percent. Specific plots consistently varied in utilization reflecting differences in production and species composition of the plots. In 1979, no meadows on the study site had signs of erosion, streambank widening or bare eroded surfaces (Kosco 1980). This was not true in 1987 (fig. 4).

1A dash indicates that the species was not found that year. Cover percentages calculated from 700 points using a 10-point frame within ungrazed cages. 2Differences in species composition are not significant (p < .05) between 1979 and 1986, except for the increase in Carex species. 3D

= decreaser, Inc = increaser, designations are from Ratliff (1985).

Inv

invader.

Category

104

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 2Production averaged approximately 3000 kg/ha prior to 1985 and 2000 kg/ha from 1985-1987. The years 1985 and 1986 produced significantly less standing crop than 1979 and 1982 which were high production years on Blodgett Forest. Mulch follows a similar pattern to production, with significantly more mulch remaining in 1980 than any other year.

Permanent Exclosure

The primary change in the permanent exclosure between 1979 and 1986 was in the statistically significant increase in Carex species from 5 to 13 percent cover. No other changes were significant, except for the decline in bare ground from 9 to .3 percent. Composition changes in the ungrazed exclosure were similar to grazed plots. Total grass cover declined from 20 to 13 percent. Rushes

slightly declined from 10 to 8 percent, while forbs remained relatively constant at 48 and 49 percent cover between 1979 and 1986. Although not statistically significant, litter increased from 8 to 15 percent and was visually apparent in the meadow. Standing crop production in 1988 was greater (2,733 kg/ha) than production on the rest of the study site (1854 kg/ha), but not significantly.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

105

Figure 3Utilization has generally increased, peaking at 82 percent in 1986. Low levels of utilization in 1980 and 1985 may be due to fewer animals on the study site in those years, or more forage available on the clearcuts.

Discussion/Conclusions
Application of the traditional range condition model based on percent composition of decreasers, increasers, and invaders would rate the current (and 1977) meadow conditions on Blodgett as fair to poor condition (Ratliff 1985). Using Crane's (1950) criteria for rating meadow condition based on the amount of bare ground or litter, the meadows on Blodgett would rate in excellent condition. Application of Forest Service (USDA 1969) soil cover criteria would rate the meadows as good. The

three models span all four possible condition ratings on the same meadows from the same data, suggesting low utility.

Species composition of the meadows on Blodgett has not changed (with the exception of an increase in Carex) in the last ten years, with or without grazing. Production has significantly declined, utilization has increased, and bare ground and litter have remained approximately the same. Thus, I conclude range condition models that link productivity to species composition are not useful for managing these types of stringer meadow systems.

106

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 4Streambank erosion and meadow trampling caused by early season use may be contributing to the decline in meadow production.

The permanent exclosure indicates that species composition did not retrace the path of range deterioration as long standing theory predicts (Heady 1975). Cattle graze the meadows on Blodgett Forest seasonlong, recently at high levels of utilization. Meadow species within reach of the livestock are grazed by the end of the grazing season, with the possible exception of the coarse, wetland rush species. One hypothesis is that the meadows were in poor condition ten years ago and remain that way today. These meadow systems have changed over the last 10 years, however. The data in the literature suggest that these meadows are not producing their potential (Ratliff and others 1983, Crane 1950, Bennett 1965). The available references suggest that these meadows should produce 5000 kg/ha on wet or moist sites, and 2400 kg/ha on dry sites. Using these criteria, peak standing crop on the wetter meadows on Blodgett is approximately 84 percent of potential. Dry meadow peak standing crop is 66 percent of potential. And moist meadow peak standing crop is only 50 percent of

potential. Low productivity may be largely due to the large component of forbs in these moist meadows, and relative lack of productive grasses and sedges. Stringer meadows are a small, but important, component of the mixed conifer ecosystem. Photographic evidence (fig. 4) indicates broken down stream banks, widened stream beds, and locally broken meadow sod in the meadows on Blodgett Forest. These factors may be contributing to low productivity, and damage to streams. Managers face a difficult problem to achieve timber management objectives through heavy livestock grazing on the clearcuts, without damage to the riparian meadow resource. More effective grazing systems to improve these meadow systems need development. Traditional range condition models are of little value either in assessing needs for improved management of these riparian systems, or evaluating the effectiveness of any newly developed practices.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

107

Acknowledgement:
I thank James W. Bartolome, University of California Berkeley; Lynn Huntsinger, California Department of Forestry and Fire Protection; and other range management graduate students from the Department of Forestry and Resource Management at Berkeley, who helped sample these meadows over the years. This study was supported by McIntyre-Stennis funds MS-2500.

Kosco, Barbara H. 1980. Combining forage and timber production in young- growth mixed conifer forest range. Berkeley: University of California: 117 p. Ph.D. Dissertation. Norusis, Marija J. 1986. SPSS/PC+ for the IBM PC/XT/ AT. Chicago, IL: SPSS, Inc. Al-H10 pp. Ratliff, Raymond D.; George, Melvin R.; McDougald, Neil K. 1983. Managing livestock grazing on meadows of California's Sierra Nevada. Davis, CA: Leaflet 21421. Coop. Extension, University of California, Division of Agriculture and Natural Resources. 9 p. Ratliff, Raymond D. 1985. Meadows in the Sierra Nevada of California: state of knowledge. Gen. Techn. Report PSW-84. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 52 p. Sampson, Arthur W. 1919. Plant succession in relation to range management. Washington, D.C.: U.S. Department of Agriculture Bulletin #791. 76 p. Sampson, A.W. 1952. Range management principles and practices. New York: John Wiley and Sons. 570 p. Society for Range Management. 1974. A glossary of terms used in range management. 2nd ed. Denver, CO: Society for Range Management. 36 p. State of California. 1988. Forest and Rangeland Resource Assessment. In review. Sacramento, CA: Supplied by California Department of Forestry and Fire Protection. Timossi, Irene. 1988. Data base output from the Wildlife Habitat Relationship data base. Berkeley, CA: Dept. Forestry and Resource Mgmt., University of California. 12 p. U.S. Department of Agriculture, Forest Service. 1969. FSH 2209.21 - Range environmental analysis handbook. San Francisco, CA: California Region (R5). Zar, J.H. 1984. Biostatistical Analysis. 2nd edition. Englewood Cliffs, N.J.: Prentice-Hall, Inc. 718 p.

References
Allen, Barbara H. 1987. Forest and meadow ecosystems in California. Rangelands 9(3): 125-128. Bennett, Peter S. 1965. An investigation of the impact of grazing on ten meadows in Sequoia and Kings Canyon National Parks. San Jose, CA: San Jose State College. 165 p. MS Thesis. Branson, F.A. 1985. Vegetation changes on western rangelands. Range Monograph #2. Denver, CO: Soc. for Range Mgmt. 76 p. Crane, Basil K. 1950. Condition and grazing capacity of wet meadows on the east slope of the Sierra Nevada Mountains. J. of Range Mgmt. 3:303-307. Dyksterhuis, E.J. 1949. Condition and management of range land based on quantitative ecology. J. of Range Mgmt. 2:104-115. Ellison, Lincoln; Bailey, Reed W. 1951. Indicators of condition and trend on high range-watersheds of the intermountain region. Ogden, UT: Agriculture Hdbk. #19. Intermountain Forest and Range Experiment Station, U.S. Department of Agriculture, 66 p. Heady, Harold F. 1975. Rangeland Management. New York: McGraw-Hill, Inc. 460 p.

108

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

AN APPLICATION OF BLM'S RIPARIAN INVENTORY PROCEDURE TO RANGELAND RIPARIAN RESOURCES IN THE KERN AND KAWEAH RIVER WATERSHEDS1
Patricia Gradek, Lawrence Saslaw and Steven Nelson 2 Abstract: The Bakersfield District of the Bureau of Land Management conducted an inventory of rangeland riparian systems using a new method developed by a Bureauwide task force to inventory, monitor and classify riparian areas. Data on vegetation composition were collected for 65 miles of streams and entered into a hierarchical vegetation classification system. Ratings of hydrologic function, vegetative structure, and vegetative use by grazing animals were employed to measure impacts by land uses and potential for recovery with proper management. The data for each of 116 stream reaches was entered into a dBase III program for tabular analyses and will be entered into the Geographic Information System (GIS) for spatial analysis and cartographic output.

Field Procedure
The first step was to examine thoroughly the results of earlier inventories. Then we developed a form to record data (fig. 1). The form includes selected parameters that had been sampled in the earlier inventories and could be resurveyed to estimate trend in conditions. In addition, it includes a combination of parameters suggested by the Bureau's Riparian Area Management Task Force (U.S. Dept. Interior 1987 a and b), Platts and others (1987). All perennial streams on public lands in the Kern and Kaweah River watersheds were inventoried. In the field, the inventory crew stratified each stream into reaches. Homogeneous reaches contained similar vegetation composition, channel morphology, degree of impact from grazing, or other management influences. A separate reach was established when any significant change in one of these three factors occurred. Most streams had several reaches since few were homogeneous enough to be considered one reach. Each reach was then marked on 7.5' scale topographic maps and on 1:24,000 scale color aerial photographs. The form in figure 1 was completed for each reach. It provided data on physical features, a riparian site function rating, and a vegetation classification for each reach. Physical Features The form records data on physical features such as landform, soil characteristics, channel substrate, channel and side slope gradient, and morphological processes. These data are used in conjunction with aerial photographs and topographic maps to generally describe the stream's character and physical constraints, and to indicate its potential for improvement in condition. The interaction between these physical features and the vegetation, animal and land use practices at a site creates complex ecosystems capable of frequent changes in conditions. Collecting this information for stratified stream reaches aids in subdividing the streams into more manageable units and identifying those critical reaches in need of special management considerations or rehabilitation.

In 1987, the Bakersfield District of the Bureau of Land Management (BLM) inventoried 65 stream miles of rangeland riparian habitats in the Kern and Kaweah River watersheds. This inventory employed state-ofthe-art methods developed by the Bureau's Riparian Area Management Field Task Force. These methods (U.S. Dept. Interior 1987a) include procedures for extensive and intensive inventories, monitoring, and vegetation classification of riparian areas. To adapt the procedure to this specific area, we used these methods as a broad guideline, and included some modifications from the work of Platts and others (1987). We found this combination of methods effective in describing, quantifying and classifying riparian ecosystems under the Bureau's management. In contrast to efforts in 1982 that centered on wildlife and water resource data needed for a grazing Environmental Impact Statement, this effort included information on vegetative structure, composition, stream morphology, and apparent trend. This work provides an overview of the structure and condition of the riparian systems in the Kern and Kaweah River Watersheds in central California. The inventory has been used to select sites for more intensive monitoring.

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. District Hydrologist, District Wildlife Biologist and Wildlife Biologist, respectively, Bakersfield District, Bureau of Land Management, U.S. Department of Interior, Bakersfield, California.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

109

Figure 1 Extensive stream riparian inventory field form 110


USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Riparian Site Function Rating The Riparian Site Function Rating (U.S. Dept. Interior 1987a) just below the center of figure 1, was the average of three ratings assigned for each reach. The three ratings, the Streambank Soil Alteration Rating, the Vegetative Bank Protection Rating, and the Subsurface Water Status Rating, evaluate interdependent factors that influence hydrologic function of the riparian site. The ratings range from "4" for excellent conditions to "1" for poor conditions. The Streambank Soil Alteration Rating (Platts and others 1983) assesses the extent of bank modification and instability caused by natural forces, livestock grazing, and other land uses (table 1). Streambank ratings are based on how far they deviate from the optimum conditions that would exist in an undisturbed state. The Vegetative Bank Protection Rating (Pfankuch 1975) was developed in recognition of the important role vegetative cover and root mats play in stabilizing streambank (table 2). In turn, stable streambanks support a greater density and variety of riparian vegetation. The vegetation protects banks from erosive forces by reducing the velocity of flood flows. Table 1 - Streambank Soil Alteration Rating (Platts and others 1983)
Rating Value Percent 4 0 1 to 25 Description Streambanks are stable and are not being altered by water flows or animals. Streambanks are stable, but are being lightly altered along the transect line. Less than 25 percent of the streambank is receiving any kind of stress, and if stress is being received, it is very light. Less than 25 percent of the streambank is false, broken down, or eroding. Streambanks are receiving only moderate alteration along the transect line. At least 50 percent of the streambank is in a natural stable condition. Less than 50 percent of the streambank is false, broken down, or eroding. False banks are rated as altered. Alteration is rated as natural, artificial, or a combination of the two.

The Subsurface Water Status Rating uses the presence of hydrophytic plant species as an indicator of whether a shallow aquifer exists (table 3). When lateral erosion or channel incision occurs, the recharging of the aquifer is impaired. The water table in the riparian aquifer is lowered, and hydrophytic species begin to decline and are replaced by upland species. Table 2 - Vegetative Bank Protection Rating (Pfankuch 1975)
Rating Value 4 Description Excellent: Trees, shrubs, grass, and forbs combined cover more than 90 percent of the ground. Openings in this nearly completed cover are small and evenly dispersed. A variety of species and age classes are represented. Growth is vigorous and reproduction of species in both the under and overstory is proceeding at a rate to insure continued ground cover conditions. A deep, dense root mat is inferred. Good: Plants cover 70 to 90 percent of the ground. Shrub species are more prevalent than trees. Openings in the tree canopy are larger than the space resulting from the loss of a single mature individual. While the growth vigor is generally good for all species, advanced reproduction may be sparse or lacking entirely. A deep root mat is not continuous and more serious erosive incursions are possible in the openings. Fair: Plant cover ranges from 50 to 70 percent. Lack of vigor is evident in some individuals and/or species. Seedling reproduction is nil. This condition ranked fair, based mostly on the percent of the area not covered by vegetation with a deep root mat potential and less on the kind of plants that make up the overstory. Poor: Less than 50 percent of the ground is covered. Trees are essentially absent. Shrubs largely exist in scattered clumps. Growth and reproduction vigor are generally poor. Root mats are discontinuous and shallow.

26 to 50

Table 3 - Subsurface Water Status Rating (U.S. Dept. Interior


1987a) Rating Value Description 4 Riparian site vegetation composition dominated by hydrophytic plants; reproduction evident. Little or no encroachment of upland plants (plants intolerant to prolonged saturated soil). Upland plants limited largely to the riparian/upland interface. 3 Riparian site vegetation composition dominated by hydrophytic plants. Some evidence of hydrophytic species decline and corresponding increase in upland plants, with upland species advancing from the riparian/upland interface. Riparian site vegetation composition is a roughly equal mix of hydrophytic and upland plant species. Upland species reproducing; little or no reproduction of hydrophytes. Water stress may be apparent in hydrophytic plants. Riparian site vegetation composition dominated by upland species, with some extending to stream channel edge. Hydrophytic species mostly scattered clumps. In extreme cases, hydrophytic species may be totally lacking. Former aquifer presence may be indicated only by isolated hydrophytic remnants such as Salix stumps, etc.

Streambanks have received major alteration along the transect line. Less than 50 percent of the streambank is in a stable condition. Over 50 percent of the streambank is false, broken down, or eroding. A false bank that may have gained some stability and cover is still rated as altered. Alteration is rated as natural, artificial, or a combination of the two. 76 to 100 Streambanks along the transect lines are severely altered. Less than 25 percent of the streambank is in a stable condition. Over 75 percent of the streambank is false, broken down, or eroding. A past damaged bank, now classified as a false bank, that has gained some stability and cover is still rated as altered. Alteration is rated as natural, artificial, or a combination of the two.

51 to 75

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

111

Figure 2 - Structure of the Vegetation Classification System (U.S. Dept. Interior 1987b)

The three ratings are summed and divided by three to obtain the mean. This value can be used to rank stream reaches by riparian site function. The reaches can be assigned a descriptive term for riparian site function using Table 4.

similarly classified locations. The approach used by the Bureau's task force was adapted from the work of Brown and others (1979, 1980) and from Paysen and others (1982). The vegetation classification data appears on the lower one-third of the form (fig. 1). The system is hierarchical such that larger, more broadly defined units are split into smaller units as one progresses down the system (fig. 2). The highest level of classification is the Vegetation Type. The two wildland vegetation types are "upland" and "wetland." Obviously all of the inventoried streams fall under the Wetland Type, so "Vegetation Type" was not listed as an entry on figure 1. The Vegetation Type is followed by the Formation. Of the six choices defined in the Bureau's system, most of our streams fell in the "Riparian Forest" or "Riparian Scrub" Formations. The Formation is followed by selection of one of four Climatic Zones. The inventoried streams fell into the "Warm temperate" or the "Cold temperate" Climatic Zones.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 4 - Riparian Site Function Rating Mean Rating Score 4 3-3.9 2-2.9 1-1.9 Rating Excellent Good Fair Poor

Vegetation Classification A system for classifying vegetation was adopted by the Bureau's Riparian Area Management Field Task Force (U.S. Dept. Interior 1987b). A vegetation classification system permits the user to group similar biotic conditions and to establish and quantify the types of habitats present. Classifying vegetation also aids when extrapolating the results of studies to other 112

Biomes are defined as natural communities characterized by a distinctive vegetation physiognomy and climatic regime within a formation (Brown and others 1979). Commonly identified Biomes were the "California Riparian Deciduous Forest and Woodland," "SierranCascade Riparian Deciduous Forest," "California Deciduous Riparian Scrub," and "Sierran-Cascade Riparian Scrub." The Biomes are followed by classification into a Series. The Series and the remaining levels of classification are based on the system of Paysen et. al. (1982). The Series is defined by the dominant species or set of species in the overstory. Dominance is determined by aerial crown cover. Examples are "Salix Series" and "Populus Series." The Association is the lowest level in the classification system and is defined as the dominant species or set of species in both the overstory layer and in subordinate layers. An example would be "Salix/Eleocharis Association". Similar Associations are aggregated into a Series, similar Series into a Biome, and so on. The category Phase is used to add further information into the classification system, primarily as a means of improving its utility for specific management applications. Phase is offset here to show that it is independent of classification level and can be used anywhere in the hierarchy. We gathered the following data under the Phase: Tree Canopy Cover, Shrub Canopy Cover, Herbaceous Cover, Grazed or Ungrazed, Percent Vegetative Use by Animals (Platts and others 1987) and Age Distribution of Woody Dominants. Mapping Each reach was mapped on topographic maps (7.5') and on 1:100,000 scale land status maps. Mapping of reaches by vegetation classification or by other parameters allows the manager to view the spatial relationship between habitat types or areas of impact. An overlay was prepared for 1:24,000 scale color aerial photographs allowing the user to pinpoint areas of interest. Reaches were identified on the overlay along with other important information such as highly impacted areas and areas of erosion. Color slides were taken of each stream reach, and additional photographs were taken of impacted areas. The data can also be used in the Bureau's Geographic Information System (GIS). Each stream is assigned a reach number and is digitized at 1:24,000 scale from the USGS topographic maps and aerial photographs. The GIS can then be used to integrate the riparian inventory information with other map themes such as grazing allotments, upland vegetation, range improvements, big game seasonal ranges, roads, etc. In addition, computer

plots can be generated to display the spatial distribution of stream reaches by vegetative classification, riparian site function rating, or other parameter. The total miles of stream for any parameter can also be calculated by GIS.

Results and Discussion


The 65 miles of inventoried streams were broken into 116 stream reaches that are fairly homogeneous in terms of vegetation composition, channel morphology and degree of impact from livestock grazing. For each of these reaches we collected baseline data on physical features, a classification of vegetation that includes general composition, structure and impacts, and a rating of hydrologic function in the riparian site function rating. The data was entered into a dBase III program that has been used to group reaches with similar conditions. Reaches can be grouped according to watershed, vegetation use by animals, erosion or water quality conditions, or by any other inventoried parameter that has been entered into the program. Table 5 lists stream reaches that are in three watersheds. The reaches in these watersheds can then be arrayed with any combination of input parameters. In this example, we also asked for the Riparian Site Function Rating (RSFRATING), whether it was grazed and the amount of vegetation use by animals (GRAZEPHASE), the Streambank Soil Alteration Rating (SSARATING) and the Vegetative Bank Protection Rating (VBPROTECT). This type of data analysis has aided in drawing conclusions and making recommendations. We have ranked the reaches by the riparian site function rating. For those with good or excellent ratings, we have developed objectives to maintain these conditions and have not proposed detailed monitoring. For those reaches with a fair to poor rating we have developed specific objectives for improvement and an intensive monitoring procedure that is being conducted at several of these sites.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

113

STREAM Caliente Ck. Caliente Ck. Caliente Ck. Berts Canyon/ Coyote Sp. Canebrake Ck. Canebrake Ck. Canebrake Ck. Canebrake Ck. Cow Canyon Ck. Cow Canyon Ck. Cow Canyon Ck. Cow Canyon Ck. Cow Canyon Ck. Spring Canyon Ck. Three Pines Ck. Three Pines Ck. Three Pines Ck. Three Pines Ck. Chimney Ck. # 1 Chimney Ck. # 1 Chimney Ck. # 1 Chimney Ck. # 2 Chimney Ck. # 2 Chimney Ck. # 3 Chimney Ck. # 4 Chimney Ck. # 4 Chimney Ck. # 4 Chimney Ck. I.T. Chimney Ck. I.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Chimney Ck. P.T. Upper Chimney Ck. Upper Chimney Ck. Upper Chimney Ck.
1Riparian 2Whether

WATERSHED Caliente Ck. Caliente Ck. Caliente Ck. Canebrake Ck. Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Canebrake Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Chimney Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck.

RSFRATING1 2.0/Fair 2.3/Fair 1.3/Poor 3.0/Good 2.7/Fair 2.0/Fair 1.7/Poor 1.3/Poor 3.0/Good 4.0/Excell. 2.0/Fair 3.3/Good 3.0/Good 1.7/Poor 1.7/Poor 2.0/Fair 2.7/Fair 2.0/Fair 4.0/Excell. 4.0/Excell. 4.0/Excell. 3.7/Good 3.3/Good 2.7/Fair 3.0/Good 4.0/Excell. 3.3/Good 2.7/Fair 3.0/Good 4.0/Excell. 4.0/Excell. 2.3/Fair 3.0/Good 1.3/Good 3.0/Good 3.0/Good 2.7/Fair 3.0/Good 3.7/Good 2.7/Fair 3.0/Good 3.0/Good 3.3/Good 3.7/Good 2.0/Fair

GRAZEPHASE2 G/Extreme G/Extreme G/Extreme U/Mod. G/Mod. G/High G/High G/Mod. G/Light G/Light G/Light G/Light G/High G/Light G/High G/Extreme G/Mod. G/Extreme U/Light U/Light U/Light G/Mod. G/High G/High G/High U/Light G/Mod. G/Light G/Mod. U/Light U/Light G/High G/Mod. G/High G/Mod. G/Light G/High U/Light U/Light U/Mod. U/Light U/Light U/Light U/Light U/Light

SSARATING3 Severe Severe Severe Major Severe Severe Severe Severe Severe Light Major Mod. Mod. Severe Severe Severe Severe Severe Light Light Light Light Mod. Mod. Major Light Mod. Mod. Mod. Light Light Mod. Mod. Major Mod. Mod. Major Mod. Mod. Major Mod. Mod. Mod. Light Severe

VBPROTECT4 Poor Good Poor Good Good Fair Fair Poor Excell. Excell. Fair Good Good Fair Poor Fair Good Fair Excell. Excell. Excell. Excell. Good Fair Good Excell. Good Fair Good Excell. Excell. Fair Good Poor Good Good Fair Good Excell. Fair Good Good Excell. Excell. Fair

#1 #2 #1 #1 #2 #2 #2 #2 #3 #3 #4 #4 #4 #5 #6

Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck. Ck.

Site Function Rating the stream was grazed and the amount of vegetative use by animals 3Streambank Soil Alteration Rating 4Vegetative Bank Protection Rating Table 5-Data from a dBase III printout of inventory data including stream reaches in three watersheds 114
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Data manipulation within the dBase III program permits us to easily quantify the extent and type of riparian systems in this area. With the vegetation classification data in particular, we are able to describe the riparian systems under our management in ways that were not possible in the past. For example, we know now that the "riparian scrub" formation is dominant on public lands in the Kern River watershed with smaller amounts of "riparian forest," "riparian marshland," and "riparian strand lands" present. By contrast, riparian areas on public lands in the Kaweah River watershed are almost entirely riparian forests. For each reach, comparisons between parameters such as "percent vegetative use by animals" and the riparian site function rating provide an indication of impacts from grazing. This inventory method and the analysis of the data in a dBase program have provided us a great amount of new information in a relatively brief period of time. The emphasis on collection of vegetative and morphological information has been an effective method of assessing the effects of livestock grazing. We recommend that this method be used in other areas in this district where this type of data is lacking. For public lands in the Kern and Kaweah River watersheds, it has documented baseline characteristics and conditions. A good baseline inventory is needed to decide where management practices should be changed and where projects should be implemented. With such data we can make better decisions on where projects should be implemented when the funds become available. The results of this inventory provide the data needed to establish management objectives and to monitor our progress toward meeting them. In addition, we have data that can be used as a starting point to assess what the appropriate levels of livestock grazing should be on these reaches. Finally, the records can be easily maintained and amended on the dBase III program. We have strongly justified the need for monitoring selected stream reaches. Maintaining an up-to-date inventory and monitoring critical reaches are required by the Federal Land Policy and Management Act of October 21, 1976 (43 U.S.C. 1701), the Public Rangelands Improvement Act of Octo-

ber 25, 1978 (92 Stat. 1803) and the Bureau's riparian area management policy. This procedure provides the method to meet these mandates and to ultimately improve the condition of riparian areas in the Bakersfield District.

REFERENCES
Brown, D.E.; Lowe, C.H.; Pase, C.P. 1979. A digitized classification system for the biotic communities of North America, with community (series) and association examples for the southwest. Journal of the Arizona-Nevada Academy of Science 14 (Suppl. 1): 1-16. Paysen, T.E.; Derby, J.A.; Conrad, C.E. 1982. A vegetation classification system for use in California: Its conceptual basis. General Tech. Rep. PSW-63. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture. Pfankuch, D.J. 1975. Stream reach inventory and channel stability evaluation. Publication number RI-75-002; Missoula, Montana: U.S. Department of Agriculture, Forest Service, Northern Region; 26 p. Platts, W.S.; Megahan, S.F.; Minshall, G.W. 1983. Methods for evaluating stream, riparian and biotic conditions. General Technical Report INT-138; Ogden, Utah: Forest Service, U.S. Department of Agriculture; 70 p. Platts, W.S.; Armor, C.; Booth, G.D.; Bryant, M.; Bufford, J.L.; Cuplin, P.; Jensen, S.; Leinkaemper, G.W.; Minshall, G.W.; Monsen, S.B.; Nelson, R.L.; Sedel, J.R.; Tuhy, J.S. 1987. Methods for evaluating riparian habitats with applications to management. General Technical Report INT-221; Ogden, Utah: U.S. Department of Agriculture, Forest Service; 177 p. U.S. Department of Interior, Bureau of Land Management, Riparian Area Management Field Task Force. 1987a. Riparian area management -inventory and monitoring; Draft Technical Reference TR# 6740-2; Denver, Colorado: 50 P. U.S. Department of Interior, Bureau of Land Management, Riparian Area Management Field Task Force. 1987b. Riparian area management - classification; Draft Technical Reference TR# 6740-3; Denver, Colorado: 44 p.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

115

THE FALLACY OF STRUCTURES AND THE FORTITUDE OF VEGETATION 1


Wayne Elmore and Robert L. Beschta2 Abstract: Given time and proper management conditions, degraded rangeland streams can often produce by natural means the same results that we expect from streambank stabilization and fisheries enhancement structures. Advantages of using vegetation and natural recovery processes include: 1) costs are likely to be lower and 2) a wide range of benefits can accrue to a recovered stream. Structures tend to lock a stream channel in place whereas vegetation allows incremental changes in channel characteristics as flow and sediment loads vary. Healthy riparian vegetation can replace itself in perpetuity, providing a resiliency which keeps banks adjusted to channels even shifting ones. Improved management of streamside vegetation, not structural additions to channels, offers the most promise for developing valuable and productive riparian systems. What Structures Do There are various reasons for installing stream structures. In arid regions, the most common reason is probably for channel stability, followed by fisheries restoration and habitat enhancement. The main goal is usually to minimize channel changes or attempt to create specific habitat features. A "stable" structure is typically considered to be indicative of a successful project. Some structures are known to aid in the deposition of sediment, slow bedload transport, store water, and retard bank erosion. Their construction has also been reported to increase fish habitat for rearing and spawning, although rigorous evaluations and monitoring of increases in biological productivity are uncommon. Even so, it is often assumed by many biologists that we have an adequate understanding of how structures can be used to modify streams and fish habitat.

State of the Art


Telling the Stream vs the Stream Telling Us A wealth of information on stream management is being presented in workshops and technical journals. Sometimes we call it habitat improvement, sometimes stream enhancement, and sometimes just stream structures. This information indicates that we have the capability to analyze, compute, design and construct projects on streams ranging in size from first to sixth order. These projects are accompanied by reporting procedures that allow us to document materials, cost per project, and "projected" benefits. Publications like "A Streambank Stabilization and Management Guide" (Commonwealth of Pennsylvania 1986) and the USDA Forest Service "Wildlife and Fisheries Habitat Improvement Handbook" (Payne and Copes 1986) take us through projects step by step. They identify project timing, size, materials and installation instructions in an attempt to ensure a successful effort. Some publications describe advantages, disadvantages, and apparent effectiveness of various structural modifications to stream channels. Diagrams and pictures provide additional information of what a structure will look like and how it is intended to operate within the stream system. The characteristics of naturally functioning stream systems can provide a basis for evaluating the relative success or failure of structural additions to streams. Figure IA provides a cross-sectional view of a degraded stream system. In this example the stream has cut down through previously deposited alluvium. As a result, the channel and associated vegetation have changed dramatically. Species typical of wetland conditions have largely disappeared and the channel continues to erode laterally. There is little subsurface storage of water and the stream is characterized by intermittent flow.

Presented at the California Riparian Systems Conference; September 22-24, 1988; University of California, Davis. State Riparian Specialist, U.S. Bureau of Land Management, Prineville, Oregon, and Hydrologist, Oregon State University, Corvallis, Oregon, respectively.

116

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 General characteristics and functions of riparian areas associated with rangeland streams.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

117

In contrast, Figure 1B illustrates a previously eroded channel that supports a diversity of riparian vegetation and has undergone recovery. The vegetation provides relative stability to streambanks and causes deposition of sediment; over time the channel has undergone aggradation. Such aggradation is often a natural consequence of allowing streamside vegetation (which may have been modified by historical grazing, logging, agriculture, or other management practices) the opportunity to again function and exert its influence on flow conditions and the characteristics of the channel. A consequence of this aggradation process is that the water table will similarly rise. In some cases, a formerly intermittent stream may flow perennially. For stream systems draining rangeland watersheds in the western United States, a change in grazing strategies is typically necessary to allow the recovery process to begin. Often, we fail to address the real problem associated with a degraded stream system, i.e., the management of streamside vegetation. We forget about the role that such vegetation plays in influencing stream dynamics and channel morphology. If most streams are currently in a degraded state due to historical management practices that have heavily impacted riparian vegetation, doesn't it seem likely that a change in management and improvement in vegetation may represent an important solution? Where improvement of a degraded channel is desired, the selection of a structural approach is seemingly driven by a desire for instantaneous gratification. We are currently using a "fast-food" approach to stream management. A certain sized pool or spawning riffle is desired and the apparent solution is to construct a structure that might immediately provide such a channel feature. Even where structural additions to a channel may help the recovery of riparian vegetation, we rarely allow several years of recovery before identifying where the structures might do the most good. Structures are almost always added to streams that are in poor ecological condition. As a result we are "telling a stream" where it needs help instead of letting the "stream tell us" where it needs help (Elmore 1987).

sills, etc.) or to structurally modify a channel with construction equipment (hydraulic excavators, backhoes, front-end loaders, etc.) will seldom provide a longterm solution to riparian problems and deficiencies. Furthermore, the construction of expensive structures allows managers to sidestep difficult management decisions (Elmore and Beschta 1987). Often, the structural "enhancement" of rangeland streams is viewed as the solution to inadequate riparian management. As a result, changes in management may not be forthcoming. Alluvial streams undergo continual channel adjustments as flow and sediment loads vary. Incremental changes in channel morphology and streamside vegetation allow such streams to withstand the wide range of dynamic forces that occur as flows fluctuate rapidly during storm runoff and snowmelt flows. An important feature of alluvial streams is that their channels will continue to adjust, change, and sometimes shift location. However, by placing permanent structures in a channel, we are attempting to lock the stream into a relatively fixed location and condition. As a result, structures that are placed in a recovering stream system are often placed where they are not needed; many are ineffective. In the rush to install expensive and often counterproductive fisheries enhancement structures, we have ignored what should be the primary management focus restoring streamside vegetation. Streamside vegetation can influence the character of a stream in a variety of ways: inputs of leaves and other plant parts as a source of energy for instream biota, shade from overstory plants, woody debris from tree species, streambank stability and cover from woody rooted species as well as sedge/grass/forb communities, and the biological processing and cycling of nutrients (Salo and Cundy 1987). Vegetation allows riparian areas to function in ways that structures cannot replicate. The resiliency provided by riparian vegetation allows these ecosystems to withstand a variety of environmental conditions. Riparian vegetation, in contrast to structures, can maintain itself in perpetuity as new plants continually replace those that die. Whereas instream structures focus narrowly at attempting to improve fish habitat, diverse and healthy riparian plant communities provide a wide range of values related to water quality and quantity, wildlife, aesthetics, channel stability and fisheries. If we are truly after productive stream and riparian ecosystems, only vegetation provides hope of a long-term solution.

Structures and Streams Many proponents of improved riparian management have reconciled themselves to spending large amounts of money to correct riparian problems. Additional funds are needed to assist in changing of streamside management, protecting sensitive stream reaches, and for additional research to better understand the complexities and management capabilities of riparian systems. Spending large amounts of money to build instream structures (e.g., gabions, dikes, check dams, rip-rap,

118

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Structures and Economics The alteration of stream channels by a wide variety of structural devices, in an attempt to enhance fish habitat or for other purposes, has gained increasing popularity in the last several years. And, based on the informa-tion presented at recent symposiums and workshops, we have become increasingly proficient at installing "stable" structures in many types of streams. However, are we judging the successes of our stream recovery enhance-ment program on the methods we use or on the results we achieve? Structures seemingly represent a quick fix and high-tech approach to degraded streams, but do they really work and are they cost effective? Bill Platts (Pers. Comm.) has indicated that structures in Big Creek (a stream with important fish habitat in northern Utah) have changed the stream into essentially a silty canal. Further, this condition now prevents the estab-lishment of willows. He is recommending the removal of instream structures along Big Creek to allow the stream to function more naturally. Based on observations in Big Creek and throughout the intermountain west, Platts has further concluded that "A dollar in stewardship is worth $10,000 in structures." Experiences in eastern Oregon and elsewhere seem to echo Platt's concerns regarding the high economic costs associated with structural modifications of stream systems. In contrast, changes in management that will benefit streamside vegetation can often be incorporated within existing management planning and action programs at little additional cost. Not only may changes in management be less costly, such actions represent a real opportunity for long-term solutions to degraded channels and riparian areas throughout much of the western United States. Vegetation recovery and responses to management, however, do require time. We cannot realistically expect all streams to recover immediately even with the most enlightened program of streamside management. However, if we want recovered riparian areas to become more prevalent in the future, we need to start today at changing our views of streamside management. We also need to reassess and reevaluate the role of structural modifications.

Conclusions
Our body requires healthy arteries and veins for nourishment. Similarly, our watersheds and streams need healthy riparian vegetation to function and provide a wide array of benefits. Now that we are able to diagnose and begin a healing process for many riparian maladies, there is little reason not to prescribe and implement treatments (Luscher Pers. Comm.). These treatments should focus on beginning the process of vegetation recovery through improved management of rangeland riparian areas; structural additions to streams are a much lower priority.

References
Commonwealth of Pennsylvania. 1986. A streambank stabilization and management guide for Pennsylvania. Commonwealth of Pennsylvania, Department of Environmental Resources, Harrisburg, PA. 79 pp. Elmore, W. 1987. Riparian management: Back to basics, the public lands during the remainder of the 20th Century. Unpublished paper presented at "The Public Lands during the Remainder of the 20th Century," 8th Annual Summer Program, Natural Resources Law Center, University of Colorado School of Law, Boulder, Colorado. June 8-10, 1987. 8 p. Elmore, W., and R.L. Beschta. 1987. Riparian areas: perceptions in management. Rangelands 9(6):260-265. Luscher, B. [1987]. State Director, U.S. Bureau of Land Management, Oregon State Office, Portland, Oregon. July, 1987. Pers. Comm. Platts, W.S. [1987]. Research Fisheries Biologist and Project Leader, Intermountain Research Station, USDA Forest Service, Boise, Idaho. December, 1987. Pers. Comm. Salo, E.O., and T.W. Cundy (Editors). 1987. Streamside management: forestry and fisheries interactions. Institute of Forest Resources, University of Washington, Seattle, Washington. 471 p. Payne, N.F., and F. Copes. 1986. Wildlife and fisheries habitat improvement handbook. Wildlife and Fisheries Administrative Report, unnumbered, USDA Forest Service, Washington, D.C. December, 1986.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

119

EVIDENCE FOR AN ALTERNATIVE LANDSCAPE CALIFORNIA ANNUAL RANGELANDS 1


Richard J. King2 Abstract: The basic tenet of annual range management in California is that perennial grasses are unable to effectively compete with the naturalized alien annual species. Evidence from the field in northern California is not consistent with this thinking. An alternative view that perennial grasses are restricted in importance from overgrazing and excessive rest is presented. Implications for range productivity and stability in riparian zones and uplands are outlined from a rancher's perspective.

POTENTIAL

IN

Roadsides and ungrazed pastures offer innumerable remnant, recovering and new stands of perennial grasses, perennial forbs, and woody species (fig. 1,3-8). Yet protection from livestock may be as detrimental to perennial grass as overgrazing. Excessive residues can accumulate within tufts and/or between plants and reduce productivity (fig. 6, 7, 9). Reproduction may suffer from reduced vigor and seed production, or an unsuitable germination or seedling microenvironment.

The goal of range managers over much of the world's grazing land is a diverse plant community dominated wherever possible by perennial plant species. Perennial vegetation including communities dominated by perennial species, with various contributions from annual species, is generally regarded as offering more stable and productive watersheds compared to sites dominated by annual species. In California annual grassland, dominance by introduced and naturalized alien annual species has become widely accepted as the new vegetation potential. Management goals for conservation and sustained productivity have shifted to a strategy of maintaining adequate residues. Most scientists and range managers believe perennials are unable to effectively compete with the aggressive annual species. This paper reports evidence from the field that potential for increasing perennial species still exists throughout northern California's annual rangelands. Implications for rangeland productivity and stability in riparian zones and uplands are outlined from a management perspective.

Management Implications
Ranchers are interested in the potential of perennials in riparian zones and uplands for a variety of reasons. Riparian zones often receive the most grazing pressure and animal impacts, yet provide critically important forage, cover, or water. The productivity and stability of many annual rangeland riparian zones often is diminished when residue on attendant upland sites is the basis for management. It is common to find accelerated gully and streambank erosion on annual range even when recommended levels of residue remain. Annuals offer little resistance to the erosive energy of concentrated flows when compared with perennials (fig. 10). Perennials, including sedges, rushes, grasses, shrubs, and trees offer far greater protection in both riparian and upland sites (fig. 7). Additionally, perennials offer other values. Managing toward perennial dominance is potentially advantageous from a manager's perspective for a wide array of reasons: a. Increased forage quality and quantity b. Longer green feed period and greater market flexibility c. Reduced erosion and sediment damage d. Improved soil structure, organic matter content, and nutrient cycles e. Improved rainfall effectiveness and forage reliability f. Greater diversity of potential ranch enterprises g. Reduced feed and supplement needs; perennial hay possible h. Reduced noxious plant and animal problems i. Reduced conflicts with environmental interests j. Improved water quality

Evidence for Perennial Potential


Evidence from personal observation that perennials have considerable potential comes from areas protected from livestock grazing and areas grazed at various intensities by livestock. Annual plant communities grazed by livestock often have a significant perennial component (fig. 1,2). Individual perennial plants are generally considered overgrazed when they are not given an opportunity to recover from defoliation. Overgrazed perennials die or exhibit reduced productivity and reproduction (fig. 1).
1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Range Conservationist, Soil Conservation Service, U. S. Department of Agriculture, Red Bluff, California. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

120

Summary and Conclusions


Field evidence throughout northern California annual rangelands suggests the potential of perennial grasses, perennial forbs, shrubs, and trees may be greater than generally believed. Overgrazing and/or excessive protection appear to be the primary factors restricting the importance of perennial grasses rather than an inability to compete with annuals.

The stability and productivity of riparian zones along water courses is often poorly protected by annual vegetation. Plant communities dominated by perennials could increase stability and productivity as well as provide numerous other benefits for rangeland resources and enterprises.

Figure 1 Highly productive hardinggrass (Phalaris Luberosa) along roadside (March). Adjacent conventionally grazed pasture has a long-established stand which is persistent despite severe grazing. Road 303, Glenn County. Rainfall zone 18-20 inches.

Figure 3 Ungrazed riparian corridor. Channel lined with trees and shrubs. Dense creeping wildrye (Elymus triticoides) and scattered trees cover flood terrace. Corral Hollow Road, San Joaquin County. Rainfall zone 1012 inches.

Figure 2 Grazed annual range dominated by needlegrass (Stipa sp.) near Flournoy, Tehama County. Rainfall zone 20-22 inches.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 4 Perennial grasses, perennial forbs, and young oaks along roadside and adjacent isolated grazed range where livestock access is restricted by banks of ephemeral stream. Vasco Road, Contra Costa County. Rainfall zone 12-14 inches. 121

Figure 5 Young oaks established in needlegrass (Stipa sp.) dominated roadside yet absent from adjacent grazed pasture. Conn Valley Road, Napa County. Rainfall zone 22-26 inches.

Figure 7 Roadside needlegrass (Stipa sp.) with individual plants hampered by accumulation of previous years' growth (January). Highway 99E, Tehama County. Rainfall zone 20-22 inches.

Figure 6 Ungrazed needlegrass (Stipa sp.) with individual plants hampered by accumulation of previous years' growth (May). Corral Hollow Road, San Joaquin County. Rainfall zone 10-12 inches.

Figure 8 Creeping wildrye (Elymus triticoides) dominating ungrazed roadside hill, sparse to absent in adjacent conventionally grazed annual range. Interstate 680 between Pleasanton and Fremont, Alameda County. Rainfall zone 18-20 inches.

122

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 9 Field of mostly dead hardinggrass (Phalaris tuberosa) plants. Plants died from excessive litter buildup within individual plants when ungrazed for many years. Livermore, Alameda County. Rainfall zone 14 inches.

Figure 10 Dense annual grassland community offers little resistance to erosive energy of overland flow (March). Johnson Road, Tehama County. Rainfall zone 20-22 inches.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

123

USE OF SUPPLEMENTAL FEEDING LOCATIONS TO MANAGE CATTLE USE ON RIPARIAN AREAS OF HARDWOOD RANGELANDS 1
Neil K. McDougald, William E. Frost, and Dennis E. Jones2 Abstract: Typical cattle use on two range units of hardwood rangeland (annual rangeland) at the San Joaquin Experimental Range, Madera County, California, left 50 percent of riparian area with less than optimum amounts of residual dry matter (RDM) for promoting seedling growth and soil protection. By relocating supplemental feeding sites away from water sources and into areas of high amounts of RDM, the impact of cattle on riparian areas was greatly reduced. In Range Unit 1, where supplemental feeding sites were relocated, only 1 percent of the riparian area was left with low amounts of RDM. In Range Unit 8, where feeding sites were left unchanged, over 50 percent of the riparian area was left with low amounts of RDM. away from riparian areas would provide more favorable conditions for forage production and soil protection than currently exist. This paper reports a study to determine if the use of riparian areas, expressed in terms of RDM remaining in the fall could be affected by a livestock management practice. Since past investigations found that relocation of salt blocks was ineffective in changing the distribution of cattle use (Wagnon 1967), we examined the relocation of the supplemental feeding areas into areas previously identified as consistently having high amounts of RDM.

Hardwood rangelands (annual rangelands) are comprised of several different land classes which vary in the amount of forage produced, time of forage growth and species composition. Distribution of use by cattle on these rangelands is highly correlated with the amount of forage production of the different land classes (Wagnon 1967); that is, the greatest amount of cattle use occurs on the highest producing areas - the riparian areas. On these rangelands the greater use of riparian areas by cattle results in lower amounts of residual dry matter (RDM) than on associated sites at the end of the grazing season (Frost and others 1988). Residual dry matter is the dry plant material left on the ground from the previous year's forage growth. Moderate amounts of RDM provide a favorable microenvironment for early seedling growth, soil protection, adequate soil organic matter and a source of low quality fall forage (Clawson and others 1982). Moderate levels of RDM for riparian areas on the San Joaquin Experimental Range have been determined to be 400 to 800 pounds per acre. During a recent 3-year supplemental feeding trial under typical feeding practices, one half of the riparian areas were consistently left with low amounts (less than 400 pounds per acre) of RDM while less productive areas were left with high amounts of RDM (McDougald and others 1987; Frost and others 1988). The low amounts of RDM do not provide the best microenvironment for seedling growth nor the best soil protection. A means of redistributing cattle use into less productive areas and

Study Area
The San Joaquin Experimental Range is located 28 miles northeast of Fresno in Madera County, California, near the center of the state and in the heart of the granite soil section of the Sierra Nevada foothills. It supports annual plant/oak woodland type vegetation and is characterized by grassy, rolling hills with a scattering of trees and occasional dense stands of brush. It is in the lower part of the woodland zone between the treeless valley floor and the higher brush and timber belts. Seeds of most herbaceous plant species germinate with the first 0.5 to 1.0 inches of fall rain. The plants grow slowly during the winter and rapidly when warm temperatures return in March. Most of the herbaceous species reach maturity in April and are mostly dry by mid-May. The climate is Mediterranean, characterized by mild, rainy winters and hot, dry summers. Annual precipitation averages 19 inches, with extremes of 9 and 37 inches. Range Units 1 and 8 were used during this trial. Both units are approximately 450 acres of which over 5 percent is considered riparian area:
Riparian Rolling, open Rolling, rocky, brushy Steep, rocky, brushy Range Unit 1 6 1 21 1 73 % 0 % Range Unit 8 7 1 5 % 63 % 25 %

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Natural Resources Specialist, University of California Cooperative Extension, Madera; Research Associate, California Agricultural Technology Institute, California State University, Fresno; and Herdsman, California State University, Fresno. The work reported here was funded in part by a Renewable Resources Extension Act Minigrant. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

124

Methods
During a 3-year study (1982-85) of range cow supplementation, the amount of RDM remaining in the fall was measured and mapped for Range Units 1 and 8 at the San Joaquin Experimental Range (Frost and others 1988; Dunbar and others in press). This 3 year period identified a pattern of use for riparian areas in those units. These patterns were used as the baseline in determining the effect of relocating supplemental feeding locations on the distribution of cattle use. Supplemental feeding sites were relocated in 1986-87 in Range Unit 1. They were placed in areas which, during the three year supplemental feeding trial, consistently had high amounts of RDM remaining in early October. Feeding locations were not changed in Range Unit 8 to provide a means of eliminating the effect of current years' weather and forage production as the cause of a shift in the distribution of cattle use. Annual precipitation and forage production were lower in 1986-87 than in the years involved in the supplemental feeding trial during which cattle use distribution was established (table 1). Use by cattle (expressed as Animal Unit Months or AUM's) was similar for all years:
Range Unit 1 1982-85 average 1986-87 449 432 Range Unit 8 469 444

riparian area. These percentages were examined to determine if the change in supplemental feeding locations in Range Unit 1 produced a change in the use of riparian areas within that unit.

Results and Discussion


Traditional supplemental feeding locations in Range Units 1 and 8 resulted in approximately 50 percent of the riparian areas being left with low amounts of RDM in early October (McDougald and others 1987)(table 2). These feeding locations were generally located in close proximity to water sources and salt locations. Cattle use in Range Unit 8 during 1986-87 followed the same general pattern as during the baseline period. During this trial and the baseline period about 55 percent of the riparian area was left with low amounts of RDM, while approximately 30 and 15 percent was left with moderate and high RDM levels, respectively (table 2). In Range Unit 1, relocating the supplemental feeding sites into areas which consistently had high amounts of RDM remaining in the fall resulted in a dramatic change in RDM levels in riparian areas. The percentage of total riparian area with low amounts of RDM was reduced from 48 percent in the 3 year baseline period to only 1 percent during this trial (table 2). The percentage of area with moderate RDM levels was also reduced from 39 percent to 27 percent while the percentage of area with high RDM amounts was increased from 13 percent to 72 percent (table 2). These results, obtained in a year of below average forage production, indicate that cattle use can be manipulated through the location of supplemental feeding sites. By moving supplemental feeding locations away from water sources and into areas where high amounts of RDM remain, the impact of cattle on riparian areas in hardwood rangelands can be greatly reduced.

(Use per year in AUM's)

In early October RDM was mapped using categories of high, moderate and low (Clawson and others 1982):
Residual dry matter (pounds per acre) Low Moderate High >800 <400 400 - 800

Table 1 Average forage production and precipitation at the San Joaquin Experimental Range, California.
Date of Production Year 1982-83 1983-84 1984-85 1986-87 54 year ave. (lbs/ac) 3,630 1,824 1,690 968 2,316 germinating rain 1 Sept. 26 Oct. 1 Oct. 17 Sept. 28 Oct. 27 Total precipitation 37.4 16.3 13.6 11.9 19.0

Table 2 Percentage of riparian area within RDM classes in early


October. Year and Unit 1982-85 Range Unit 1 Range Unit 8 1986-87 Range Unit 1 Range Unit 8 Low 48 59 1 54 Moderate 39 29 27 33 High 13 12 72 13

Amounts of RDM were determined by the Comparative Yield Method (Haydock and Shaw 1975) and visual estimation. The acreages within each RDM class were mapped, measured and expressed as percentages of total 125

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

References
Clawson, W. James, McDougald, Neil K. and Duncan, Don A. 1982. Guidelines for residue management on annual range. Cooperative Extension, Division of Agricultural Sciences, University of California. Leaflet 21327. 3 p. Dunbar, John R., McDougald, Neil K., Jenkins, Audrey S., Daley, Dave A., Topping, Bart L. and Frost, William E. 1988. Range cow supplementation. California Agriculture 42(5):8-10 Frost, William E., McDougald, Neil K., and Clawson, W. James. 1988. Residue mapping and pasture use records for monitoring California annual rangelands. University of California Cooperative Extension. Range Science Report No. 17. 9 p.

Haydock, K.P. and Shaw, N.H. 1975. The comparative yield method for estimating dry matter yield of pasture. Australian Journal of Experimental Agriculture and Animal Husbandry. 15:663-670. McDougald, Neil K., Clawson, W. James, Frost, William E., and Duncan, Don A. 1987. Residue mapping and pasture use records for monitoring California annual rangelands. Abstracts, 40th Annual Meeting, Society for Range Management. No. 51. Wagnon, Kenneth A. 1967. Use of different classes of range land by cattle. California Agricultural Experiment Station Bulletin 838. Division of Agricultural Sciences, University of California. 16 p.

126

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

CLARK CANYON (MONO COUNTY) RIPARIAN DEMONSTRATION AREA1


John W. Key and Mark A. Gish2 Abstract: The Clark Canyon riparian demonstration area was established in 1984 within the East Walker River subbasin of Mono County, California. Destabilization of the meadow sections of the stream and the upper stream reaches contributed to an increase of suspended sediments, turbidity, and stream channel widening in the lower stream reaches where a viable population of rainbow trout is found. Several different treatments have been implemented to (1) restore meadow riparian areas to high levels of productivity, (2) stabilize active erosion and gully development (headcutting), (3) improve aquatic habitat from poor to good condition, and (4) improve wildlife cover and downstream fish habitat. These treatments include changes in grazing management practices and the construction of several types of instream structures. Management and Planning (CRMP) process and serves as a demonstration area for several different techniques for riparian area rehabilitation.

The Clark Canyon Riparian Demonstration Area


The Clark Canyon Riparian Demonstration Area is located about 5 kilometers east of the town of Bridgeport, Mono County, California. The climate of the area is cold and semi-arid. Annual precipitation ranges from 20 to 40 centimeters and occurs mainly as snow in the winter. The elevation is 2200 to 2260 meters. The soils on the canyon bottom are formed in mixed stratified alluvium. They are dark colored, deep, moderately fine textured and poorly to somewhat poorly drained along the drainage bottoms; and dark colored, deep, moderately coarse to fine textured, and well drained along the drain ways and side slopes. Soils on the canyon sides are rocky and shallow to moderately deep, with moderately coarse textures over moderately fine textured subsoils. Clark Canyon Creek covers a total of 6.4 stream kilometers within the East Walker River subbasin. It is a perennial stream which receives much of its flow from subsurface water as it flows through the canyon. Stream flow is fairly constant through the summer months, and at the junction with Aurora Canyon Creek it is as cool or cooler than the upper reaches. In 1979 an intensive stream survey recorded an average stream width of 0.9 meters and an average stream depth of 3 centimeters and a discharge rate of 0.0057 cubic meters per second (cms). Constituent water analysis from stream samples taken at the time of the 1979 stream survey revealed undesirably high levels of iron (Fe) and manganese (Mn). Also, heavy algal growth was reported in the meadow sections of the stream, attributed to livestock trailing and heavy grazing use. An increase in suspended sediments (turbidity) was also reported in the lower stream reaches where a viable rainbow trout population occurs. The four major vegetation types in Clark Canyon (Barbour and Major 1977, Ratliff 1985) are: The montane meadow vegetation type is located on the moister alluvial sites. The ecological grouping of major meadow species in the riparian demonstration

In 1984, the Bishop Resource Area of the Bureau of Land Management (BLM) established the Clark Canyon erosion control project. This project implemented several different treatments to restore riparian meadow areas and to stabilize active erosion and gully development. Riparian meadow areas are unique and among the most productive and important ecosystems on the public lands. They display a greater diversity of plant and wildlife species and vegetation structure than adjoining ecosystems (Bureau of Land Management 1987). Riparian meadows are narrow, highly productive plant communities located along streams. These meadows are usually dominated by grasses and grass-like plants with shrubs often as a major vegetative component. Livestock preference on riparian meadows has been reported as a major influence on overall grazing distribution on mountain rangeland (Gillen and others 1985, Platts 1986). Livestock are attracted to these areas for water, shade, and vegetation that remains green after upland forage has dried out. Riparian watershed values include water table recharge, soil erosion reduction, flood water control, and sediment and nutrient collection (Thomas 1986). Surface flooding and elevated water tables have been reported as having a definite influence on plant vigor of certain shrub species (Ganskopp 1986). This erosion control project has become a successful example of riparian area management using the Coordinated Resource

Presented at the California Riparian Systems Conference, September 22-24, 1988, Davis, California. Soil Scientist and Range Conservationist, Bureau of Land Management, U.S. Department of the Interior, Bakersfield, California, and Bishop, California, respectively.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

127

area is as follows:
Scientific Name (Munz and Keck 1973, Messick 1982) Primary Meadow Species Carex lanuginosa C. nebraskensis C. rostrata Deschampsia caespitosa Secondary Meadow Species Aster adscendens Eleocharis pauciflora Hesperochiron californicus Hordeum brachyantherum Muhlenbergia asperifolia M. richardsonis Poa cusickii P. nevadensis Ranunculus cymbalaria Senecio hydrophilus Invader Meadow Species Iris missouriensis Common Name

Woolly Sedge Nebraska Sedge Beaked Sedge Baltic Rush Long-leaf Aster Common spikerush Centaur Meadow Barley Alkali Muhly Mat Muhly Cusick Bluegrass Nevada Bluegrass Desert Buttercup Swamp Groundsel Western Iris

areas were on a downward trend with streambanks being impacted by heavy live-stock use. This condition was causing the elimination of natural overhanging banks and concomitant widening of the stream. The stream profile was changing to wide, shallow, and saucer shaped; water was becoming warmer due to les s shade. Erosion problems were evident with active gullying at the upper reaches of the perennial flow. Shallow gullying was occurring throughout the riparian meadow areas and large headcuts (active erosion and gully development) were rapidly lowering the water table in two of the moist meadows along the stream by draining them. Plans to rehabilitate the riparian areas in Clark Canyon were started in 1982. A combination of several treatments including grazing management were proposed to (1) restore meadow riparian areas to high levels of productivity, (2) stabilize headcutting, (3) improve aquatic habitat from poor to good condition, and (4) improve wildlife cover and downstream fish habitat. Representatives from the Bureau of Land Management, Forest Service, and California Department of Fish and Game visited the area and assisted with the preliminary plan development to repair stream channel damage caused by livestock grazing. The complete erosion control project included improved grazing management and the construction of gully control structures. Improved grazing management would restrict grazing until gully development has been stabilized. The gully control structures were designed to fill gullies, elevate water tables, and control further erosion until natural vegetation becomes vigorous enough to become permanent. A series of instream structures were planned for construction to control active erosion and gullying (Key 1987). Two major headcuts were planned to be stabilized to control further cutting. The upper 2.4 kilometers of Clark Canyon were divided into different stations and site-specific recommendations developed for each station. The watershed erosion control project consisted of two large wire mesh structures, three small gabion basket structures, three single fence rock-check dams, four double fence rock-check dams, and one loose rock headcut treatment. In 1984 the project was started at the headwaters of Clark Canyon Creek with the goal of contouring the steep side slopes for natural revegetation, raising the water table in the riparian meadow portions, and preventing further degradation of the remaining riparian meadows. The two large wire mesh structures were constructed in 1984 near the head-waters of the stream, where the gullying was the most severe (fig. 1). These two structures were lined with an erosion filter fabric to trap fine soil particles behind the structures (fig. 2).

This meadow type is dominated by Nebraska sedge, Common spikerush, and bluegrasses. Vegetative cover is more than 85 percent and vegetative production is high. The riparian meadow vegetation is found in stringers along the creek. Several small bogs with peat moss (Sphagnum fimbriatum) are found at major spring sources of Clark Canyon Creek within the project area. The Great Basin sagebrush vegetation type is located on the drier alluvial sites. Vegetative cover was inventoried at about 30 percent and vegetative production was low. Wyoming big sagebrush (Artemisia tridentata wyomingensis) is the dominant species with 35 to 90 percent composition. The pinyon-juniper vegetation type is located in shallow, moderately coarse textured soils on the stony canyon sides and the rocky uplands surrounding Clark Canyon. One leaf pinyon pine (Pinus monophylla) and Utah juniper (Juniperus osteosperma) dominate this vegetative type. The deciduous woodland vegetation type is located along perennial streams in the canyon. It is dominated by a dense growth of narrow-leaved willow (Salix exigua) and arroyo willow (S. lasiolepis). These willows are accompanied by a large undergrowth of wood rose (Rosa woodsii) and, where willows adjoin a meadow, squaw current (Ribes cereum) is found.

Erosion Control and Rehabilitation


The 1979 intensive stream survey of Clark Canyon rated aquatic habit at from fair to poor condition. Nonmeadow riparian areas rated poor due to the heavy livestock damage. Streambanks were badly trampled and denuded of vegetation; willows and other streamside vegetation were severely hedged. Riparian meadow 128

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 1 Site of large wire mesh structure, Clark Canyon, California, June 22, 1983.

Figure 2 Large wire mesh structure covered with an erosion control filter fabric, Clark Canyon, California, July 24, 1984.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

129

The three smaller gabion basket structures were constructed in 1984 in shallow gullies in the meadow area. Initially, they were not lined with erosion control filter fabric (fig. 3). In 1985, the upper gabion basket was modified and lined with erosion control filter fabric to improve its effectiveness (fig. 4). The three single fence rock-check dams were installed in October 1985 with hand labor, erosion control filter fabric, woven wire fence, and metal fence posts (fig. 5). These small instream structures have been very successful (fig. 6) in elevating the water table and controlling further erosion. The four double fence rock-check dams were installed in 1986 and 1987, and have proven to be successful in both controlling further erosion and elevating the water table. Proper spacing between structures depends upon the gradient of the gully and stream. The minimum interval used had the crest of one structure level with the apron of the structure above it. Key locations for the structures are immediately below a junction of two or more small gullies, at narrow points of the gully, and at points where the gully is not eroding rapidly. The loose rock headcut treatment was implemented in 1986. The head wall was cut back so rock could be placed easily under the sod crest. Foundation rock was placed at the toe of the loose rock rip-rap to ensure that the rip-rap would not slide away from the headcut. Rock was carefully placed and hammered into the head wall soil to ensure close contact. Rip-rap was built under the lip of the sod crest and over the foundation rock to form an apron. Flat wedge shaped rock was used and hammered into and under the sod crest. Water was allowed to flow over the sod crest across the rock riprap, across the apron, and downstream in the stream channel.

Initiating action in the Clark Canyon area was complicated by the fact that Clark Canyon was an unfenced grazing allotment boundary between two different grazing allotments (Aurora Canyon and Travertine Hills). As a result, two CRMP activities were startedone for each grazing allotment. Major issues to be resolved specific to Clark Canyon were (1) evidence of accelerated erosion in Clark Canyon, (2) poor vegetation conditions in meadows and other riparian areas in Clark Canyon, and (3) livestock grazing use conflicts caused by using Clark Canyon as the grazing allotment boundary between two grazing allotments and the impact of historic livestock trailing through Clark Canyon. In order to resolve livestock grazing impacts in Clark Canyon, BLM personnel from the Bishop Resource Area in January 1984 consulted with the livestock permittees affected by the plan. Concerns focused on availability of water and forage for livestock, livestock trailing through the canyon, and livestock drift between the Travertine Hills and Aurora Canyon allotments. Written consensus was reached for the following: Fencing and Livestock WaterTwo fence projects were implemented to create watering gaps for livestock. The upstream water gap fence was constructed between the exclosures surrounding the two large wire mesh structures. These exclosures were constructed to protect the recovering streambanks from livestock trampling. Approximately 36 meters of stream between the exclosures were left accessible for livestock watering. The second water gap fence was constructed approximately 0.8 kilometers downstream near the public-private land boundary. This fence controls cattle drift along an easily accessible portion of Clark Canyon and provides a water gap at a commonly used draw west of the creek. It also provides a means to confine cattle within the 0.8 kilometer intensive management area. Both fences are inspected and maintained annually by BLM personnel. Livestock TrailingSheep trailing through the intensive management area was continued with the stipulation that no bedding, grazing, or watering occur, and that all allotment boundary gates are kept closed. One sheep permittee on a neighboring allotment traditionally trails two bands of sheep through the project area during the first week of June on the way to his allotment and in mid-October on his return to his home ranch. The sheep permittee in the Travertine Hills allotment does not use Clark Canyon for trailing. Cattle trailing through the project area is not authorized.

Coordinated Resource Management and Planning


Because of the involvement of both public and private land in the project, Coordinated Resource Management and Planning (CRMP) was used to prepare a resource plan for the public land grazing allotments and associated private lands. CRMP is a resource planning process used to address resource problems, based upon a philosophy that resource conflicts can best be solved at the local level by direct communication among all interested groups and individuals. It is based on the premise that people who meet together voluntarily will find common ground as they interact with one another and have a chance to observe resource problems firsthand on the ground (Nevada Coordinated Resource Management and Planning Task Group 1983).

130

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 3 - Unlined gabion basket structure, Clark Canyon, California, July 28, 1984.

Figure 4 - Same gabion basket structure as in Figure 3 after modification and lining, Clark Canyon, California, July 17, 1986.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

131

Figure 5 Installation of single fence rock-check dam, Clark Canyon, California, October 10, 1985.

Figure 6 Single fence rock-check dam, Clark Canyon, California, August 4, 1987. 132
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Stocking LevelThe grazing capacity within the intensive management area was determined by calculating the animal-unit months (AUM) based upon the 1980 forage inventory. Twenty AUM were determined available for annual use during the period of June 15 through October 31. The AUM authorized allows only a few cattle to use the area season long or not at all. The normal use by each permittee is as follows:
AUM Season of Use 10 cows 6/16 - 8/15 6 cows 6/16 - 9/30

Plots 2 and 3 were permanently obliterated by deposited sediments after spring runoff in 1986 and were not reestablished. Plot 4 was partially obliterated in 1987. An initial stream profile utilizing sag tape transects was conducted at the time the two large wire mesh structures were constructed. A total of 10 sag tape transects (fig. 7) were made annually since 1984 to document changes in stream profile and sediment deposition collection behind each of the two large wire mesh structures. The transects followed procedures described by Ray and Megahan (1979). Stream stability indicators were evaluated initially in 1984 at the two large wire mesh structures and on an annual basis since that time. The results of this rating is used in conjunction with the stream profile and sediment deposition measurements and revegetation monitoring to determine stream condition and trend. Weather information collected at the Bodie State Park is used to evaluate the influence of annual weather patterns on the other components being monitored. Results of all the monitoring studies are evaluated each fall to determine progress towards stated objectives.

Travertine Hills cattle permittee: Aurora Canyon cattle permittee:

The cattle numbers are generally pairs (the cow and her calf). The grazing plan allows the Travertine Hills permittee to use the area for two consecutive seasons of use followed by three consecutive seasons of use by the Aurora Canyon permittee. Utilization levels at season's end approach the high end of the moderate class, for herbaceous vegetation, using the Key Forage Plant Method. The moderate class is described as: "The rangeland appears entirely covered uniformly as natural features and facilities will allow. From 15 to 25 percent of the number of current seed stalks of key herbaceous species remain intact. No more than 10 percent of the number of herbaceous forage plants are utilized."

Monitoring
The following methods were used to monitor the success of the erosion control project: trend and photo plots, stream profile and sediment deposition, stream stability indicators, and weather. A total of four trend plots were established within the project area to monitor grazing impactsone inside the two exclosures and two outside the exclosures (one upstream of the second large wire mesh structure and one downstream of the second large wire mesh structure). Each plot was 0.9 meters by 0.9 meters square and was placed on the streambank close to the water's edge. The plot locations were also selected on the basis of having some existing vegetation in place as opposed to purely bare ground. The methodology used was the Photo Plot Method which involves taking an overhead photo of the plot and a panoramic photo of the background from a located photo point marker. Plot readings involve determining species identification, the number of mature plants or seedlings by species, and estimating the number of 1/16 units per square foot (0.0929 square meters) each plant species and litter occupies within each plot (table 1). These data are then multiplied by a factor for plot size to determine plot totals.
Table 1. Three years of vegetation monitoring data
1985 1986 1988 73.1 60.8 0 22.4 Obl.

Plot 1 ( Inside Upper Exclosure) Percent Composition 95.4 96.0 Percent Cover 3.7 33.4 Number of Seedings 7.0 7.0 Percent Litter 0.2 1.4 Plot 2 (Outside Upper Exclosure) Percent Composition Percent Cover Number of Seedings 87.7 0.5 24.0 Obl.

Percent Litter 0.1 Plot 3 (Inside Lower Exclosure) Percent Composition Percent Cover Number of Seedings 53.1 7.9 3.0 Obl. Obl.

Percent Litter 7.0 Plot 4 (Outside Lower Exclosure) Percent Composition Percent Cover Number of Seedings Percent Litter 94.6 16.5 7.0 0.9 95.3 17.6 26.0 1.3 Obl.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

133

References
Barbour, Michael G.; Major, Jack. 1977. Terrestrial vegetation of California. New York: John Wiley and Sons; 1002 p. Bureau of Land Management. 1987. Riparian Area Management. TR 175-814. Washington, D.C.: U.S. Department of the Interior; 4 p. Ganskopp, David C. 1986. Tolerance of sagebrush, rabbitbrush, and greasewood to elevated water tables. Journal of Range Management 39(4): 334-337. Gillen, R.L.; Krueger, W.C.; Miller, R.F. 1985. Cattle use of riparian meadows in the Blue Mountains of Northeastern Oregon. Journal of Range Management 38(3): 205-209. Key, John W. 1987. Small instream structure construction for meadow restoration in Clark Canyon, California. In: Proceedings of the California Watershed Management Conference; 1986 November 18-20; West Sacramento, California. University of California, Wildlands Resource Center; Report No. 11:161. Messick, Tim. 1982. The flora and phytogeography of the Bodie Hills of Mono County, California and Mineral County, Nevada. Arcata, California: Humboldt State University; 167 p. M.S. Thesis. Munz, Philip A.; Keck, David D. 1973. A California flora with supplement. Berkeley: University of California Press; 1681 p.; supplement 224 p. Nevada Coordinated Resource Management and Planning Task Group. 1983. Nevada coordinated resource management and planning handbook. 2nd edition. 15p. Platts, William S. 1986. Managing riparian stream habitats. In: Proceedings of Wyoming water and streamside zone conference; 1986 April 28-30; Casper, WY. 59-62. Ratliff, Raymond A. 1985. Meadows in the Sierra Nevada of California: state of knowledge. Gen. Tech. Rep. PSW84. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 52 p. Ray, G.A.; Megahan, W.F. 1979. Measuring cross sections using a sag tape: a generalized procedure. Gen. Tech. Rep. INT-47. Odgen, UT: Intermountain Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 12 p. Thomas, Allan E. 1986. Riparian protection/enhancement in Idaho. Rangelands 8(5): 224-227.

Figure 7 Sag tape transect, Clark Canyon, California, July 17, 1986.

Acknowledgements
We thank Terry Russi and Larry Saslaw, wildlife biologists, and Mark Blakeslee and Patricia Gradek, hydrologists, Bureau of Land Management, U.S. Department of the Interior; and Tom Felando, forest hydrologist, Inyo National Forest, Forest Service, U.S. Department of Agriculture, for their efforts and support in implementing this project. We also thank Lori Key for her assistance in editing and manuscript preparation. This project was initially funded by the State of California Energy Resource Fund (ERF Project 7120 5661 1028) managed by the California Department of Fish and Game.

134

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

SOUTHWESTERN WOODY RIPARIAN VEGETATION AND SUCCESSION: AN EVOLUTIONARY APPROACH 1


R. Roy Johnson, Peter S. Bennett, and Lois T. Haight Abstract: Interrelationships between flooding and climax woody vegetation in riparian ecosystems of the desert Southwest are discussed. The lack of succession in woody desert upland and desert riparian plant communities results from opposite stresses, the former from aridity, the latter from flooding. Today's "wet riparian big five" are northern tree species of hydroriparian and mesoriparian (wet riparian) ecosystems; remnants of the Arcto-Tertiary Geoflora. The "dry riparian big five" are tree or subtree constituents of xeroriparian ecosystems occurring as Madro-Tertiary remnants at the northern extremes of their ranges. Human activities have interrupted normal flood regimes of Southwest rivers, resulting in desertification and endangering native riverine ecosystems.
2

Changes in the diagnostic vegetative structure of a given riparian ecosystem is an early warning of broader problems. Closer examination invariably shows triggering processessoil erosion or deposition, dewatering of the system resulting in the changing of perennial or intermittent streams to ephemeral watercourses, and other signs of riparian and aquatic degradation. In order to better interpret and understand the full implications of such historic and continuing changes one must understand the driving forces of riparian ecosystems. The interrelationships of flooding, succession (or lack thereof) and different woody plant regimesArctoTertiary vs. Madro-Tertiary Geoflorasneed to be fully understood.

Woody Riparian Vegetation


Trees and shrubs are major components of riparian ecosystems. Their elevated woody structure forms the characteristic landscape feature which visually distinguishes riparian ecosystems from their surroundings, especially in the Southwest deserts. It is also this structure which provides habitat for the highest concentrations of birds in North America; provides shade for recreationists, fish, and cattle; and forms the biomass of the basic trophic level in these productive riparian ecosystems (Johnson and Carothers 1982).

Riparian Big Five Lowe (1961, 1964) developed the concept of the riparian "big-five" in reference to five widespread riparian trees in the Arizona lowlands: cottonwood (Populus fremontii), willow (Salix bonplandiana) and others, e.g. (S. gooddingii), sycamore (Platanus racemosa wrightii), ash (Fraxinus velutina), and walnut (Juglans microcarpa major). Other riparian species of the desert Southwest show more limited geographic and elevational distributions (fig. 1). These or closely related species also occur throughout the California and Southwest desert lowlands in general (table 1).

Table 1 Lowe's (1961, 1964) "wet riparian big five" or Arcto-Tertiary Geoflora riparian species of the desert Southwest with California (Holstein 1984) and eastern United States analogs (Little 1980). Common Name Southwest Ash Cottonwood California Eastern United States F. pennsylvanica P. sargetii P. d. P. racemosa J. hindsii S. g. P. occidentalis J. nigra S. n.

Fraxinus velutina F. latifolia [F. v. subsp. pennsylvanica] Populus fremontii [P. deltoides & subsp. P. palmeri, P. wislizenii] P. f.

Sycamore Walnut Willow

Platanus racemosa var. wrightii Juglans microcarpa var. major Salix goodingii [S. nigra subsp. goodingii]

1 Presented 2

at the California Riparian Systems Symposium; September 22-24, 1988; Davis, California.

Senior Research Scientist, Research Scientist, and Research Assistant, respectively; Cooperative National Park Resources Studies unit, University of Arizona, Tucson.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

135

ran Desert of the United States and Mexico but is too frost-tender to live in either of the colder Mohave or Chihuahuan deserts. Carothers and others (1974) were the first to quantify the importance of the wet riparian big five to avian populations. Parallel wildlife importance has been recently discussed for xeroriparian ecosystems, of which the dry riparian big five are major components (Johnson and Haight 1985).

Plant Succession
Plant succession is the replacement of a plant community on a given site by other plant communities on the same site over a period of time, usually occurring in a predictable order. Although European botanists were the first to record the processes, two mid-western botanists, Fredrick E. Clements (1916) and his student, J.E. Weaver (Weaver and Clements 1938), further developed the concept while working in the eastern United States. Odum (1969) identified three general characteristics of plant succession. It is an orderly, predictable process with one natural community modifying the environment and thereby allowing for the establishment of subsequent communities (seres). This culminates in a stabilized, energy-optimizing ecosystem (climax community). Absence of Desert Riparian and Upland Succession Figure 1 Elevational and ecological amplitudes (spans) for the "wet riparian big five" and "dry riparian big five" in the desert Southwest. Other species, e.g. Arizona alder (Alnus oblongifolia) were excluded from these categories by Lowe (1961, 1964) and Johnson and Lowe (1985) due to their narrow amplitudes (after Benson and Darrow 1981; Kearney and Peebles 1969; Little 1980). These generally occur as dominant woody species of hydroriparian (perennial) and mesoriparian (intermittent) ecosystems (Johnson and Lowe 1985) along lowland watercourses. A more technical paleobotanical term is "Arcto-Tertiary riparian big five." Using the same criteria for xeroriparian systems along desert washes Johnson and Lowe (1985) listed the "dry riparian big four" mesquite (Prosopis spp.), catclaw acacia (Acacia greggii), ironwood (Olneya tesota), and blue paloverde (Cercidium floridum); we here add a fifth species, desert willow (Chilopsis linearis), to complete the "dry-tropic riparian big five." These five species occur in the Lower Colorado subdivision of the Sonoran Desert of southern California. As with the "wet riparian big five," these or closely related species are widely distributed in the Southwest deserts. The one exception is Olneya, which is widespread throughout the SonoThe development of plant communities in Southwest deserts differs markedly in several ways from that in classical, mesic vegetation. Clementsian succession, so well studied and documented in the eastern United States, does not occur either in desert upland or riparian communities. This was discussed at length by Lowe (1959) and for upland systems first pointed out by Shreve (1951) who wrote: "It is not possible to use the term climax with reference to desert vegetation ...If a particular community is destroyed without change in the soil, the earliest stage in the return of vegetation will be the appearance of young plants of the former dominants. Not only do the same species reappear at the outset, but their first individuals ultimately constitute the restored community." Shreve (1951) did not differentiate between upland and riparian habitats in his writings about desert vegetation. Lowe (1964) was the first to specifically address riparian succession in the Southwest, finding a lack of succession for Southwest desert riparian ecosystems where cottonwoods and willows are both "pioneer" and "climax" riparian species (in Clementsian terms). Further, the same individual plant is both a pioneer and a member of the climax vegetation.

136

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Although succession does not occur in either riparian or upland ecosystems of the desert Southwest two different, opposite stresses are involved. In riparian ecosystems flooding ("too much" water) is the driving function behind plant community development. In upland systems aridity ("too little" water) limits community development, excluding succession. Interestingly, Lowe (1964) also points out the lack of succession in a high elevation spruce-fir forest in northern Arizona. Here the stressful environment results from water occurring only as ice much of the year, thus being unavailable for uptake by plants. Consequently plant communities at high elevation and/or high latitude may demonstrate some of the same xeric characteristics of desert upland plant communities, including abbreviated or non-existent successional stages. In a paper on "riparian succession" Campbell and Green (1968) suggested "mosaics of various seral stages" but did not demonstrate succession for Southwest riparian vegetation. They stated, "The channel vegetation probably never reaches a climax hierarchy due to periodic flood disturbances such as erosion, inundation, and deposition." Reichenbacher (1984) likewise discussed riparian succession in the Southwest without substantiating it. He also considered the establishment of different riparian species under different physical conditions at different microsites as successionalstating, "On the convex bank early seres are initiated by seedling establishment while mature vegetation on the concave bank is undercut." Rather than succession, he was discussing the aforementioned vegetational mosaic mentioned by Campbell and Green (1968) characteristic of riparian vegetation. This patchiness in riparian vegetation is largely a result of shifting channels and abandoned meanders. Riparian communities are composed of species as they occur along a moisture gradient. Reichenbacher (1984) and Campbell and Green (1968) actually made a better case for the continuum concept, a concept especially discussed by Whittaker (1975). A continuum is "a gradient of environmental characteristics or of change in the composition of communities" (Ricklefs 1979). This concept postulates that plant communities are composed of a collection of species, each with a specific gradient and coexisting not because of biological affinities for one another but because of common environmental needs. The continuum concept has been discussed for other areas; e.g. for riverine aquatic ecosystems and their physical and biological parameters as expressed by headwater to mouth gradients (Vannote and others 1980). Similarly, riparian continua have been discussed along drier to wetter moisture gradients both for headwater to mouth (intrariparian) and upland to deepwater (transriparian) gradients (Johnson and Lowe 1985).

Mechanisms preventing either riparian or upland succession in the southwest have not been thoroughly examined, but for riparian ecosystems there are contributing factors. Competition is an important factor in plant succession (Odum 1971; Ricklefs 1979). Shreve (1951) pointed out that in the southwestern deserts the struggle is with the environment rather than inter- or intraspecific competition with other plants, and "the frequency on the desert of extensive communities which are simple in composition is not due to the poverty of the perennial flora so much as to the severity of the physical conditions." Compared to desert upland systems, Southwest riparian ecosystems are much more mesic and thus might be expected to show a closer similarity to eastern deciduous hardwood forests in respect to competition and succession. This, however, is not the case. The different plant communities occurring along a river that are interpreted by some as different seral stages occur side by side in different physical habitats. For example, mesquite bosques often occur on higher terraces or alluvial fans while adjacent floodplain bottomlands are vegetated by cottonwood-willow forests. Mesquites, because of their long tap roots, can secure water from deeper water tables than cottonwoods and willows. Therefore, because of differing abilities to obtain water, the two communities occur side by side at the same time rather than one community following the other on same site. Two related processes to consider in the lack of riparian succession in the Southwest are catastrophic floods and the long period of time required for successional stages (seres) to evolve into a climax community. Our examination of historic photos show flooding as a major reason for riparian plant communities in the Southwest not reaching more than 100 years in age before being greatly disrupted or destroyed. These photos, taken at chronological intervals along southwest rivers, show groves of mature cottonwoods and mesquite bosques at different sites along a river from one decade to the next. Periodic floods scour the substratum and destroy attendant groves at a given locality. Subsiding flood waters then create conditions for a seedbed, allowing germination of a new stand of young cottonwoods either in the same spot or elsewhere (Brady and others 1985). Thus, the relatively short intervals between catastrophic floods in the Southwest deserts does not allow sufficient time for succession. Factors that have changed flow regimes along southwest lowland streams, greatly affecting development of riparian plant communities, are discussed later under floods.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

137

Although individual sandbars, terraces, or riparian plants may be short-lived, the riparian ecosystem is stable. The most pertinent statement in literature is by Lowe (1964:62): The southwestern riparian woodland formation is characterized by a complex of trees, and their plant and animal associates, restricted to the major drainageways that transgress the landscape of desert upward into forest. It is incorrect to regard this biotic formation as merely a temporary unstable, seral community. It is an evolutionary entity with an enduring stability equivalent to that of the landscape drainageways which form its physical habitat. That is, it is a distinctive climax biotic community.

increasingly serious environmental degradation (Betencourt 1988; Cooke and Reeves 1976; Dobyns 1981; Rea 1983; and Turner 1988). Other human activities that have interfered with flow regimes, especially water storage and diversion projects, are discussed by Johnson and Carothers (1982). Overgrazing and consequent desertification has also greatly affected natural riparian community development (Brown and others. 1977; Dobyns 1981; Glinski 1977; Johnson and Simpson 1988; Reichenbacker 1984). Without concentrated efforts to reverse many of these disastrous, often unnecessary, activities, desertification of riparian and aquatic ecosystems of the Southwest will irreversibly proceed, thereby destroying some of our most valuable natural resources.

Acknowledgements Desertification of Southwest Riparian Ecosystems


Many historically recorded Southwest perennial and intermittent streams are now ephemeral watercourses. Analyses of factors involved in the disastrous conversions of these watercourses have been conducted by investigators such as Hastings (1959), Hastings and Turner (1965), Cooke and Reeves (1976), Dobyns (1981), and Betancourt and Turner (1988). In many cases there has been an almost total loss of both the aquatic and riparian resources that originally attracted European settlers to these linear desert oases. We thank Dr. Charles H. Lowe, University of Arizona; Dr. Steven W. Carothers, SWCA, Inc, Environmental Consultants; and Mr. James M. Simpson, Phoenix, Arizona, for numerous discussions about the contents of this paper. Mrs. Donna Marchant assisted with the typing and technical writing aspects.

References
Benson, L.; Darrow, R.A. 1981. Trees and shrubs of the Southwestern Deserts. Tucson: Univ. of Ariz. Press; 416 p. Betancourt, J.L.; Turner, R.M. 1988. Historic arroyocutting and subsequent channel changes at the Congress Street Crossing, Santa Cruz River, Tucson, Arizona. In: Whitehead, E.E.; Hutchinson, C.F.; Timmermann, B.M.; Varady, R.G., eds. Arid lands: today and tomorrow. Boulder, Colo: Westview Press; 1353-1372. Brady, W.; Patton, D.R.; Paxson, J. 1985. The development of Southwestern riparian gallery forests. In: Johnson, R.R.; Ziebell, C.D.; Patton, D.R.; Ffolloitt, P.F.; Hamre, R.H., tech. coords. Riparian ecosystems and their management: reconciling conflicting uses. Ft. Collins, Colo: USDA For. Serv. Gen. Tech. Rep. RM-120; 39-43. Brown, D.E.; Lowe, C.H.; Hausler, J.F. 1977. Southwestern riparian communities: their biotic importance and management in Arizona. In: Johnson, R.R.; Jones, D.A., tech. coords. Importance, preservation and management of riparian habitat: a symposium. Ft. Collins, Colo: USDA For. Serv. Gen. Tech. Rep. RM-43.; 201-211 Campbell, C.J.; Green, W. 1968. Perpetual succession of stream-channel vegetation in a semiarid region. J. Ariz. Acad. Sci. 5:86-98. Carothers, S.W.; Johnson, R.R.; Aitchison, S.W. 1974. Population structure and social organization in southwestern riparian birds. Amer. Zool. 14:97-108. Clements, F.E. 1916. Plant succession: analysis of the development of vegetation. Washington: Publ. Carnegie

Before the North American import of Old World cattle and the invention of the steel plow, most of the soils of the United States were largely covered by vegetation. The root-filled sod served as a sponge, allowing water to run off slowly toward drainages and clean water to percolate into the watertable. Runoff water from the uplands was partially cleansed by riparian processing (biological and physico-chemical processes of riparian ecosystems) before entering branching, meandering networks of drainages and flowing slowly downstream. When floods overtopped the riverbanks, rich alluvial soils were deposited, increasing the fertility and productivity of riparian ecosystems. Today these channels are carrying increased volumes of debris and silt-laden water from tributaries that have been greatly altered by humans, e.g. storm drains and gullies, from uplands denuded of absorptive vegetation and topsoil through misuse of the land by livestock and humans. Many of these channels have been heavily silted-in or, more often at lower desert elevations, have become incised and denuded, allowing

138

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Inst.; 242:1-512. Cooke, R.U.; Reeves, R.W. 1976. Arroyos and environmental change in the American Southwest. Oxford, England: Oxford Res. Stud. in Geog., Clarendon Press; 213 p. Dobyns, H.F. 1981. From fire to flood: historic human destruction of Sonoran Desert riverine oases. Ballena Press Anthro. Pap. No. 20. San Rafael, CA: Ballena Press; 222 p. [out of print] Glinski, R.L. 1977. Regeneration and distribution of sycamore and cottonwood trees along Sonoita Creek, Santa Cruz County, Arizona. In: Johnson, R.R.; Jones, D.A., tech. coords. Importance, preservation and management of riparian habitat: a symposium. Ft. Collins, Colo: USDA For. Serv. Gen. Tech. Rep. RM-43.; 116-123. Hastings, J.R. 1959. Vegetation change and arroyo cutting in southeastern Arizona. J. of the Ariz. Aca. of Sci. I:60-67. Hastings, J.R.; Turner, R.M. 1965. The changing mile. Tucson: Univ. of Ariz. Press; 317 p. Holstein, G. 1984. California riparian forests: deciduous islands in an evergreen sea. In: Warner, R.E.; Hendrix, K.M., eds. California riparian systems. Berkeley: Univ. of Calif. Press; 2-22. Johnson, R.R.; Haight, L.T. 1985. Avian use of xeroriparian ecosystems in the North American warm deserts. In: Johnson, R.R.; Ziebell, C.D.; Patton, D.R.; Ffolloitt, P.F.; Hamre, R.H., tech. coords. Riparian ecosystems and their management: reconciling conflicting uses. Ft. Collins, Colo: USDA For. Serv. Gen. Tech. Rep. RM-120; 156160. Johnson, R.R.; Lowe, C.H. 1985. On the development of riparian ecology. In: Johnson, R.R.; Ziebell, C.D.; Patton, D.R.; Ffolloitt, P.F.; Hamre, R.H., tech. coords. Riparian ecosystems and their management: reconciling conflicting uses. Ft. Collins, Colo: USDA For. Serv. Gen. Tech. Rep. RM-120; 112-116. Johnson, R.R.; Simpson, J.M. 1988. Desertification of wet riparian ecosystems in arid regions of the North American Southwest. In: Whitehead, E.E.; Hutchinson, C.F.; Timmermann, B.M.; Varady, R.G., eds. Arid lands: today and tomorrow. Boulder, Colo: Westview Press; 1383-1393.

Johnson, R.R.; Carothers, S.W. 1982. Riparian habitat and recreation: interrelationships and impacts in the Southwest and Rocky Mountain region. Eisenhower Consortium Bull. 12. Ft. Collins, Colo: Rocky Mt. For. and Range Exper. Sta., USDA For. Serv.: 31 p. Kearney, T.H.; Peebles, R.H. 1969. Arizona flora. Berkeley and Los Angeles: Univ. of Calif. Press; 1085 p. Little, E.L. 1981. The Audubon Society field guide to North American trees: eastern region. New York: Alfred A. Knopf; 714 p. Lowe, C.H. 1959. Contemporary biota of the Sonoran Desert: problems. In: Univ. of Ariz., Arid Lands colloquia, 19581959:54-74. Lowe, C.H. 1961. Biotic communities in the sub-Mogollon region of the inland Southwest. J. Ariz. Acad. Sci. 2(1):40-49. Lowe, C.H. 1964. Arizona's natural environment. Tucson: Univ. of Ariz. Press; 136 p. Odum, E.P. 1969. The strategy of ecosystem development. Science 164:262- 270. Odum, E.P. 1971. Fundamentals of ecology. Philadelphia, Pa.: W.B. Saunders Co. 574 p. 3rd ed.

Rea, A.M. 1983. Once a river: bird life and habitat changes on the middle Gila. Tucson: Univ. of Ariz. Press; 285 p. Reichenbacher, R.W. 1984. Ecology and evolution of southwestern riparian plant communities. In: Desert Plants 6(1):1-62. Ricklefs, R.E. 1979. Ecology. 2nd ed. New York: Chiron Press, Inc.; 966 p. Shreve, R. 1951. Vegetation and flora of the Sonoran Desert: vol. I, vegetation. Carnegie Inst. Washington Publ. 591:1-192. Vannote, R.L.; Minshall, G.W.; Cummins, K.W.; Sedell, J.R.; Cushing, C.E. 1980. The river continuum concept. Can. J. Fish. Aquat. Sci. 37:130-137. Weaver, J.E.; Clements, F.E. 1938. Plant ecology. New York: McGraw-Hill Book Co. Whittaker, R.H. 1975. Communities and ecosystems. 2nd ed. New York: MacMillan Publ.; 385 p.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

139

RELATIVE NATURE OF CONSIDERATIONS 1

WETLANDS:

RIPARIAN

AND

VEGETATIONAL

Peter S. Bennett, Michael R. Kunzmann, and R. Roy Johnson2 Abstract: Riparian ecosystems have been divided into three basic types; hydroriparian, mesoriparian, and xeroriparian associated in the desert lowlands with perennial, intermittent, and ephemeral water respectively. Floral species associated with these ecosystems can be divided into four categories; obligate, preferential, facultative, and nonriparian. Additionally, various site characteristics such as (1) latitude, (2) elevation, (3) soil, (4) slope, (5) exposure, (6) water periodicity, and (7) water chemistry, create a complex matrix of terms with each one needing delineation. Few of these parameters can be readily measured in the field. We examine the classifying of vegetation according to the species presence (actual or potential) and the evolutionary/developmental history of vegetational communities and implications in wetland and riparian land classification. differences are particularly pronounced when comparing wetlands and riparian lands of the desert Southwest to the more mesic East. Even though eastern derived definitions have generally prevailed for national programs, more than half of the nation's wetlands and riparian lands are in the more arid West.

Discussion
Johnson and others (1984a) introduced the "xeroriparian," "mesoriparian," and "hydroriparian" ecosystem descriptors to reflect ecological differences, especially for ecosystems associated with ephemeral, intermittent, and perennial waters in the Southwest deserts. Floral species associated with these three ecosystems can be divided into four categories (the 4-Category System): obligate, preferential, facultative and nonriparian (Johnson and others 1984a). There is no classification system, let alone rules and regulations, that address the combination of these seven entities. The current regulations and the proposed seven day test for wetland inundation (40 CFR Section 230.3 and EPA Wetland Identification and Delineation Manual) strictly applied in the west, would include only the wettest environments and exclude many of the periodic (ephemeral and perhaps intermittent) western wetlands and riparian lands (Lowe and others 1986). By using these seven descriptors (three riparian and four floral) we are talking about wetland/riparian land ecosystem expression interpreted from "the top down" including botanical, hydrologic and soil parameters. However, plants are the ultimate synthesizers of environmental information, and vegetation composition alone may be considered the key to (1) latitude, (2) elevation, (3) soils, (4) slope, (5) exposure, (6) water periodicity, (7) water chemistry, (8) evapotranspiration rate and many other factors. Wetland/riparian land classification and delineation is best accomplished by examination of factors that are always present and measurable. For example, consider the situation at Wilcox Playa, a large lakebed in the Chihuahuan Desert grassland of Southeastern Arizona. The playa surface is quite barren, with salty soils at its lowest elevation, becoming occupied by various salttolerant plants at its periphery and finally by mesquite

Most scientists believe that they have a workable definition of "wetlands" and "riparian" in an ecological context. Writers of laws and regulations have this same confidence. Resource managers and administrators also believe they have equally usable definitions. For each term, one can find dozens of definitions in dictionaries, manuals, textbooks, and other sources. Some of these definitions are similar, others are very different. EPA regulations (40 CFR Section 230.3) state "... areas that are inundated or saturated with surface or ground water at a frequency and duration sufficient to support, and that under normal conditions do support, a prevalence of vegetation typically adapted for life in saturated soil conditions" are wetlands. Under idealized conditions, these common sense, commonly held views are adequate to the task of definition. Each person defining these terms uses concepts appropriate to their own particular situation and background. At best each definition generally applies to a local, regional, or particular ecosystem situation. However, application of these local concepts presents severe problems when these definitions are applied at a national scale. Johnson and Carothers (1982), Lowe and others (1986), Johnson and others (in press) and Kunzmann and others (in press) present information demonstrating that there are generally significant structural and functional differences between eastern and western wetlands. These
1Presented 2

at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

Research Scientist, Ecologist and Senior Research Ecologist, respectively, United States National Park Service, Tucson, Arizona. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

140

subtrees up-slope. The lower elevation playa surface is intermittently flooded by a thin sheet of water, and buried to a depth of more than a meter at least every century. Mostly its surface is dry. If we are to follow the Environmental Protection Agency's 7 day saturation criteria to classify or map the extent of wetlands on the playa, we would have to observe the extent and persistence of the pool after each rain. But this does not draw the line for the 100 year (1 meter + depth) event. While it is true that soil saturation can often be judged by soil characteristics, the relationship of the 7 day saturation rule to soil characteristics has not been well established. In fact, 7 days of inundation, or other defined intervals, may or may not support "a prevalence of vegetation typically adopted for life in saturated soil conditions" (quotes from 40 CFR Section 230.3). Where the substratum is composed of coarse sand or cobbles typical hydric soil characteristics are not present. Other problems include judging soil saturation in unusual western situations, e.g. where the fine textured playa surface lies beneath a layer of shifting wind-blown sand. Additional problems include determining the bases for classifying wetlands/riparian lands where permanent water lies well below the soil surface but still within reach of phreatophytic species, such as the mesquites found at Wilcox Playa. Many similar problems can arise regarding classification by using criteria, e.g. hydrologic regimes, that are not always present and visible. There is a strong similarity in visual contract between riparian ecosystems and those of the desert uplands of the Southwest, whether mesoriparian hydroriparian, or xeroriparian (Johnson and others 1984a). Many of the latter might not be delineated as wetlands under classification systems such as Cowardin and others. (1979) or U.S. Army Corps of Engineers and E.P.A. Guidelines. However, these xeroriparian plant communities are generally characterized by a higher visible differentiation from the vegetation of adjacent uplands, even than most riparian ecosystems, e.g. the visual differences between bottomland hardwoods along running streams and deciduous hardwood forests of the eastern U.S. In addition to the visual contract between xeroriparian ecosystems and adjacent uplands, xeroriparian communities are generally composed of measurably different biotas from those of adjacent communities. Thus biotic components differentially utilize xeroriparian ecosystems whether classifiers considers them wetlands or not. Johnson and others (1984a) determined that vegetational differences do occur between difference or occur between different order washes as well as between uplands and xero-riparian lands. The preferential use of xeroriparian habitats compared to desert upland habitats has been quantified for birds. Johnson and Haight (1985, 1988) found up to 10 times (or

greater) the number of individuals and species of birds along desert washes in comparison to adjacent uplands in the Chihuahuan and three subdivisions of the Sonoran Desert. We propose addition of a visual comparison component to the wetland/riparian land definition; that vegetated wetlands/riparian lands be defined as "those whose vegetation is dominated by obligate, preferential or facultative wetland/riparian plants whose species occurrence, or woody plant density, or stature differ by more than 25 percent (an arbitrary figure) compared to the surrounding non-wetland communities." After additional analysis the 25 percentage difference may be modified to a more realistic figure if necessary to reflect the actual vegetation pattern. Now classification can be based on something that is visible and measurable at any time. The requirement that vegetation be composed of wetland/riparian plants would prevent misclassification of areas where effective soil moisture differences in upland situations result in vastly different vegetation types, e.g. along ridge-lines, where slope/aspect control available moisture and vegetational differences greater than 25 percent are commonplace, between north and south facing slopes, for example. In general, the 25 (or whatever) percent rule would apply to wetland/riparian obligate or preferential species. It is not true, however, for facultative species, e.g. in the desert Southwest. Foothills paloverde (Cercidium microphyllum) is a widely distributed, diagnostic upland species in the Arizona Upland Series (Brown and others 1979). It grows on rocky hillsides with saguaro (Cereus giganteus), desert ironwood (Olneya tesota), mesquite (Prosopis spp.), and triangleleaf bursage (Ambrosia deltoidea) along the eastern and northern limits of the Sonoran Desert (annual precipitation 7-13 in.). Near Yuma, Arizona, (annual precipitation approximately 3 inches) these species still occur together but are strictly confined to washes where they behave as obligate xeroriparian riparian species. Redbud (Cercis occidentalis), an eastern upland species, typically occurs along water courses in the arid west. In xeric environments creosotebush (Larrea tridentata) grows to a stature of a meter or less and plants are spatially separated by 2 meters or more. Yet along 1st or 2nd order desert washes (Johnson and others 1984b) it occurs as a xeroriparian plant with individuals occasionally 3 to 4 meters tall and separated by less than a meter. Our proposed 25 percent difference in species composition, woody plant density, or plant height, would correctly classify each of these examples above as wetlands/riparian lands in relation to their surrounding communities.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

141

Conclusions
These wetland/riparian land classification and identification difficulties will continue so long as ecologists persist in differentiating between uplands and wetlands/riparian lands on bases other than plant species growing in the natural communities. A paloverdetriangleleaf bursage community should be classified as such regardless of where it is growing. Differentiation into a "wetland" category should be in relation to surrounding ecosystems. Although on the surface this proposal seems inapplicable to legal absolutes, its ecological flexibility allows the definition to be adopted to varying local conditions and/or habitats. Until a scientific consensus is reached about the classification and demarcation of wetlands/riparian lands, we suggest that the legal system depend more heavily on the opinions of expert wetland/riparian ecologists, than on the rule-making ability of attorneys and administrators.

Consortium Bull. 12. Ft. Collins, CO: Rocky Mt. For. and Range Exper. Sta., USDA For. Serv.: 31 p. Johnson, R.R.; Carothers, S.W.; Simpson, J.M. 1984b. A riparian classification system. In: Warner, R.E. and Hendrix, K.M., eds. California riparian systems; Berkeley, CA: Univ. of California Press; 375-380. Johnson, R.R., and L.T. Haight. 1985. Avian use of xeroriparian ecosystems in the North American warm deserts. In R.R. Johnson, C.D. Ziebell, D.R. Patton, P.F. Ffolloitt, and R.H. Hamre, tech cords. Riparian ecosystems and their management: reconciling conflicting uses. USDA For. Service Gen. Tech. Rep. RM-120. Ft. Collins, CO; 156-160. Johnson, R.R., and L.T. Haight. 1988. Avian use of xeroriparian systems in the Big Bend Region, Texas. Contr. Pap. of the Sec. Symp. on Resources of the Chihuahuan Desert Region: United States and Mexico, No. 16. Chihuahuan Desert Res. Inst., Alpine, TX 9 p. Johnson, R.R.; Warren, P.L.; Anderson, L.S.; Lowe, C.H. 1984a. Stream order in ephemeral watercourses: a preliminary analysis from the Sonoran Desert. Hydrology and water resources in Arizona and the Southwest. In: Proc. of the American Water Resources Assoc, Ariz. section. Tucson, AZ: Univ. of Arizona. 14:89-100. Kunzmann, M.R.; Bennett, P.S.; Johnson, R.R. In Press. Riparian and wetland considerations in a national riparian program: an urban perspective. In: Urban wetlands and riparian habitat: Proc. of a national symposium. Berne, NY: Assn. State Wetland Managers. Lowe, C.H. 1964. Arizona landscapes and habitats. In: The vertebrates of Arizona. Univ. of Ariz. Press, Tucson. 270 p. Lowe, C.H.; Johnson, R.R.; Bennett, P.S. 1986. Riparian lands are wetlands: the problem of applying eastern American concepts and criteria to environments in the North American Southwest. In: Hydrology and water resources in Arizona and the Southwest. Proc. of the American Water Resources Assn, Ariz. section. Tucson, AZ: Univ. of Arizona; 14 :89-100.

References
Brown, D.E.; Lowe, C.H.; Pase, C.P. 1979. A digitized classification system for the biotic communities of North America, with community (Series) and association examples for the Southwest. J. Ariz.-Nev. Acad. Sci. 14 (Suppl. 1):116. Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. USDI Fish and Wildlife Service FWS/OBS-79/31, Washington D.C. 103 pp. Johnson, R.R.; Bennett, P.S.; Haight, L.T. Southwestern riparian vegetation and succession: an evolutionary approach. (These Proceedings.) Johnson, R.R.; Carothers, S.W. 1982. Riparian habitat and recreation: interrelationships and impacts in the Southwest and Rocky Mountain region. Eisenhower

142

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

THE RIPARIANNESS OF A DESERT HERPETOFAUNA1


Charles H. Lowe2 Abstract: Within the Mojave, Sonoran, and Chihuahuan Desert subdivisions of the North American Desert in the U.S., more than half of 143 total amphibian and reptilian species perform as riparian and/or wetland taxa. For the reptiles, but not the amphibians, there is a significant inverse relationship between riparianness (obligate through preferential and facultative to nonriparian) and desertness. In addition to the nondesert species (N=36) present, there are two evolutionary kinds of desert species in the herpetofauna: true desert species (N20), and desert-included species (N=87); the former are obligate specialists, the latter are facultative generalists. Quantitative aspects of desertness, riparianness, species richness, nondesert taxa and others are examined. (1942, 1951): Chihuahuan Desert, Sonoran Desert, Mojave Desert, Great Basin Desert. Creosotebush (Larrea divaricata) is one of the abundant dominants absent from the Great Basin Desert, sometimes referred to as a "cold desert." 'The three major subdivisions in which creosotebush is among the major dominants are referred to here and elsewhere as "warm deserts." These are the Chihuahuan, Sonora, and Mojave Deserts. In this preliminary report the data for the areas of these three major desert subdivisions that occur in the U.S. sector are extracted from a database that includes the amphibians and reptiles from the larger area of all four subdivisions of the North American Desert. Thus the present report contains in addition to all of the Mojave Desert, only those parts of the Chihuahuan Desert and the Sonoran Desert that lie north of the U.S.A. Mexico international boundary; most of the geographic area of the Chihuahuan and Sonoran Deserts lie in Mexico. In the present workinvolving less than the entire North American Desert herpetofaunathe database involves a total of 12 potential R on D combinations for a total of 143 taxa requiring direct field observation; that yields a potential number greater than 1700 for determinations. There are certain combinations for R and D that are "invalid" or "error" combinations, a subject treated in a longer (subsequent) report on this subject for the North American Desert herpetofauna inclusive of the United States and Mexico. The Semidesert Grassland.In the international borderlands of the study area, the North American desertgrassland, or semidesert grassland, lies adjacent to the warm deserts. It is sometimes not understood that the desert-grassland is grassland, not desert. Present deserts evolved out of grasslands and scrublands during the late Tertiary (Axelrod, 1950, 1958, 1979). Throughout the desert areas of the world, biologists make a correct distinction between the desert proper, i.e., desertscrub and its adjoining grassland, steppe, or scrubland. In our Southwest, the Holocene dry-tropic grasslands, called desert-grassland (Shreve 1917) or semidesert grassland (Little, 1950), obviously represent grassland environments. While the present quantitative analysis has been cognizant of desert environment versus nondesert environment, and of desert species versus nondesert species, the difficult cases of the mosaics and ecotones of desertscrub versus semidesert grassland, especially in the Chihuahuan Desert arena, provide a particularly difficult challenge for ecologic and biogeographic analysis.

A large part of the herpetofauna of North America is located extensively and abundantly in riparian habitats. No other terrestrial vertebrate group is a better indicator of the biological health of riparian ecosystems. Within the "warm deserts" of the Southwest United States more than half of the total amphibian and reptilian species perform as riparian and/or wetland taxa. Riparian taxa are obligate, preferential, or facultative components of riparian ecosystems. Thus including the nonriparian condition, four levels of riparianness (R), or riparian dependency, are recognized (Dick-Peddie and Hubbard 1977, Johnson, and others 1987). Moreover, for deserts, in addition to the distinction between desert species and nondesert species, there is a clear distinction between two evolutionary kinds of desert species: true desert species and desert-included species. True desert species are obligate specialists in the real sense that they have evolved within desert environments, while the desert- included species tend to be facultative generalists that include desert environments in their much wider and often widely extensive ecological and geographical distributions. Thus including the nondesert condition, three levels of desertness (D) are recognized (Lowe 1968; and others, 1986):
Desertness Nondesert (ND) Desert (D) Desert-Included (DI) Obligate Desert (DT) Riparianness Obligate (RO) Preferential (RP) Facultative (RF) Nonriparian (NR)

The North American Desert.There are four major subdivisions in the North American Desert of Shreve

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Department of Ecology and Evolutionary Biology University of Arizona, Tucson, Arizona 85721

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

143

There is a recent regional tendency to use the synonym semidesert grassland, recommended by the U.S. Forest Service. The synonym desert- grassland also continues to be used, as it has been for most of this century. Recent Reports.During recent years there has been a sharp increase in field research directed to amphibian and reptilian populations in Southwest riparian ecosystems. [In the manuscript editing process, a brief review of these papers was eliminated by the Station editors to "reduce non-relevant reference listings."]

Methods
The areas in the North American Desert for the present analysis, i.e., the study area, lie outside of the Great Basin Desert. Specifically they are the parts of the Mojave, Sonoran, and Chihuahuan Deserts that lie in the six United States of California, Nevada, Utah, Arizona, New Mexico, and Texas. The scheduled degree of desertness (desert to nondesert) and riparianness (obligate to nonriparian) was determined for each amphibian and reptilian taxon in the study area. These ecological determinations for the taxa are provided from direct field observations, mostly those of the author in the southwestern United States and northern Mexico. Over several years colleagues have been generous in both verifying and extending the database for some of the species in and between both California and Texas, which are the longitudinally limiting U.S. states in the desert coverage reported. These field experts are Robert L. Bezy, George L. Bradley, Charles J. Cole, F. R. Gehlbach, S. F. Hale, Peter A. Holm, Howard E. Lawler, Brent E. Martin, Hugh K. McCrystal, Philip C. Rosen, C. R. Schwalbe, Robert C. Stebbins, Thomas R. Van Devender, John W. Wright, and Richard G. Zweifel. It is worth noting in this connection that there are two sources of commonly used information that are inadequate for this type of fauna- wide ecological analysis. These are the field guides for, and the museum specimens of, the amphibians and reptiles. These two sources often have been abused as substitutes for lack of worker-knowledge in ecological and biogeographical investigations requiring ecological input for reports in both the primary and nonprimary literature. The obvious reason that field guides in particular cannot provide the detailed ecological information required is the fact that the modern field guides and similar books were never intended for this type of service and obviously are not designed for providing it. Among the several existing field guides for amphibians and reptiles in North America, with regard to treatment of the natural biotic communities in which the species live as well as in its illustration and other excellence, by far the best in the West is by Stebbins (1985).

Because in both theory and practice riparianlands (characterized in important part by imported water) fall within the broader concept of "wetlands," there is some current confusion in their application to plant and animal taxa occurring on natural western landscapes. Part of the problem lies in the often strong and confusing seasonal periodicity of both in-situ water (wetlands) and imported water (riparianlands) in desert environments. Table 1 indicates an expected concordance in the concepts and thus the schedules for wetland species status (developed in the eastern U.S.) and riparian species status. This schedule from obligate riparian to nonriparian (Table 1) was used in the present investigation. The regression model (Fig. 1) employed for the testing reported here for the herpetofauna involves a streamlining of riparian and wetland terminology. For example, the spadefoot toads (genus Scaphiopus) perform as both obligate wetland and obligate riparian corridor species (WO/RO) in arid and semiarid environments. In the present study such are scored with a single value for OBLIGATE, whether the taxon is primarily or wholly RO or WO, or RO/WO or WO/RO. Moreover, for further simplification as in modeling, RO is used to represent the obligate position on the riparian/wetland ordinate. An essential notation is that a taxon whether plant or animal is an obligate riparian one when it is directly dependent on the riparian system during any phase of its life cycle.

Figure 1 Inverse relationship of riparianness on desertness in a desert herpetofauna. Equations in Table 4.

144

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 1 Comparative schedules for riparian land and wetland plant species. Wetland schedule after Reed (1986). Riparian schedule modified after Dick-Peddie and Hubbard (1977), and Johnson and others (1987). Frequencies follow wetlands convention (see Reed 1986).

Riparianlands Species Status Obligate Riparian (RO)1

Frequency (%)

Wetlands Species Status

>99

Obligate (OBL) Always found in wetlands under natural (not planted) conditions; may persist in nonwetlands if planted there by man or in wetlands that have been drained, filled, or otherwise transformed into nonwetlands

Capable of natural establishment only in the riparian environment

Preferential Riparian (RP)2 More frequently in the riparian environment than in the adjoining upland Facultative (RF)2 Subequally in the riparian environment and the adjoining upland. Nonriparian (NR)2 = Upland Upland taxon, present or absent in the riparian environment

99-67 Facultative Wetland (FACW) Usually found in wetlands; occasionally found in nonwetlands

66-33

Facultative (FAC) Sometimes found in wetlands; also occurs in nonwetlands

<33

Facultative Upland (FACU) Seldom found in wetlands and usually occurs in nonwetlands

1. Frequency of occurrence--in wetland/riparian versus nonwetland/nonriparian-across the entire distribution of the species. 2. Potentially obligate riparian (RO) locally, as in driest sector of species distribution.
145

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Results and Discussion


Table 2 indicates the extent and species richness of the herpetofauna. In the total of 143 species in the study area, reptiles outnumber amphibians approximately 4 to 1. Species of lizards and snakes are present in subequal numbers. Separately, the lizards and the snakes each outnumber the turtles by 5 to 1. With lizards and snakes taken collectively (Squamata), the squamate-turtle ratio is approximately 10 to 1. The North American Desert herpetofauna is, of course, somewhat richer than indicated in Table 2, which represents a smaller total desert area. The final form and content of this table and Tables 3 and 4 await finalization regarding desertness and riparianness for certain taxa, and the actual evolutionary (species) status of others that remain unsettled. For example, to what extent are the genera Gambelia (Leopard Lizards) and Uma (Fringe-toed Lizards) polytypic? Gambelia is treated here as a monotypic genus, and Uma as a polytypic genus with one species in the Mohave-Sonoran desert arena. The introduced error swing between maximum splitting and maximum lumping is about 2%. While none of the few remaining systematic and ecologic decisions referred to above will affect significantly the overall conclusions drawn from the data set, all of them are of much interest for the completeness as well as correctness of the data set. N Amphibia %

Tables 3 and 4 indicate the degree of desertness in the herpetofauna and its correlation with riparianness. The data for the reptiles (Table 4) fit the regression model (Fig. 1) that predicts an inverse relationship between desertness (X) and riparianness (Y). Ultimate adaptations underlying the strong negative correlations seen in TaDT (N) Amphibia (N = 27) Salamanders Frogs species sums % of total 27 desert vs. nondesert Reptilia (N = 116) Turtles Lizards Snakes species sums % of total 116 desert vs. nondesert Herpetofauna (N = 143) species sums 20 14.0% 87 60.8% 75% 36 25.2% 25% 11 7 19 16.4% 1 1 34 34 69 59.5% 76% 9 9 10 28 24.1% 24% 0 1 1 3.7% 1 17 18 66.7% 70% 1 7 8 29.6% 30% DI (N) ND (N)

Salamanders Frogs and Toads Total Amphibians Amphibian % of Herpetofauna Reptilia Turtles Lizards Snakes Total Reptiles. Reptilian % of Herpetofauna Total Herpetofauna

2 25 27

7.4 92.6

% of total 143 desert vs. nondesert

100.0 18.9 Table 3 Number (N) and percent of totals for true desert (obligated desert) species (DT), desert-included (facultative desert) species (DI), and non-desert species(ND). See Table 1.

11 54 51 116

9.5 46.5 44.0 100.0 81.1


taxa (species) all reptiles turtles lizards snakes df regression equation 114 R = 3.566 - 0.962 D 9 R = 4.130 - 1.174 D 52 R = 3.270 - 0.814 D 49 R = 3.487 - 0.958 D r P

-0.763 <.001 -0.938 <.001 -0.677 <.001 -0.744 <.001 -0.707 <.001

143

100.0
lizards and snakes 103 R = 3.362 - 0.875 D

Table 2 Number (N) and percent of totals for species in the combined herpetofauna within the Mohave, Sonoran, and Chihuahuan subdivisions in the United States sector of the North American Desert.

Table 4 Regression equations for riparianness (R) on


desertness (D) for reptilian groups collectively within the Mohave, Sonoran, and Chihuahuan sub-divisions of the U.S. sector of the North American Desert. See Fig. 1.

146

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

ble 4 are the taxa-specific water balance ratios and the reptilian amniotic egg. Within the boundaries set by those constraints, trophic and other energy-behavioral adaptations drive the habitat-selection exhibited. The data for amphibians do not fit the inverse R on D model for reptiles. Amphibians as a group in a desert environment are (entirely to essentially) obligate, and thus the degree of riparianness (Y) is for the most part statistically independent of degree of desertness (X). As indicated in Fig. 2, the R on D regression coefficients for amphibian taxa in desert environments are predicted to be not significantly different from zero (b 0). In natural habitats wherever they are, the primary reproductive wedlock of the anamniotic amphibians to water and water-wetness is overriding.

the 3-desert study area. A combined 105 species of lizards and snakes comprise 90% (90.5) of the reptilian species in the study area, with approximately half of the 90% contributed by each systematic group46.5% lizards, 44.0% snakes. Turtle species (N = 11) comprise the remaining approximate 10% (9.5) of the reptilian total. A strong inverse relationship between desertness (D) and riparianness (R) emerges for the three major reptilian groups in the North American Desert-turtles, (Fig. 2) lizards, and snakes. While this relationship was predicted from the model, the correlation is stronger than expected, especially when all groups are combined (r=-0.763,P < .001). In the ecological data on desertness (D) and riparianness (R) for lizards and snakes, a strong pattern similarity is seen for them in the desert area treated; the sums for the lizards and snakes are nearly identical, and neither the slopes nor the intercepts are significantly different.

Summary and Conclusions


Desertness.Nondesert species, and desert species of two evolutionary types, characterize the desert herpetofauna. The two kinds of desert species are (1) true desert species which are obligate specialists, and (2) desertincluded species which are facultative generalists; the two are non-equivalent desert species. In a herpetofaunal total of 143 species in the North American Desert study area, a respectable 25% (25.2) are nondesert species (ND). In the 75% (74.8) that are desert species (D), 61% are desert-included species (DI), and 14% are true (obligate) desert species (DT)a small but not unexpectedly small percentage of true desert species in the North American Desert. Riparianness.The descriptor Riparian is used in this report as a generic term inclusive of both the riparian and the more purely wetland situations; riparian (and wetland) species are either obligate, preferential, or facultative, and the nonriparian species are upland species. Approximately 40% (41.2) of the 143 total species in the herpetofauna are nonriparian (NR). Nearly 60% (58.8) are riparian and/or wetland species and exhibit various degrees of compensation for moisture, with more than half of the 60% (37.1%) restricted to the obligate riparian (RO) and/or obligate wetland (WO) ecological position. Relative Species Richness.There are approximately 4 reptilian species to every 1 amphibian species in the North American "warm desert" ecosystems investigated. Similarly , turtles are outdistanced about 5 to 1 by both the lizards and the snakes. The lizards and snakes are virtually equal in species richness. Reptiles.A total of 116 reptilian species comprise approximately 80% (81.1) of the total herpetofauna in
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 2. The relationship of riparianness and desertness in amphibian and reptilian taxa in the North American Desert herpetofauna.

Amphibians.A total of 27 amphibian species comprise approximately 20% (18.9) of the herpetofauna of the desert study area. All but two of the 27 are frogs and toads. The other two are salamanders, the Tiger Salamander (Ambystoma tigrinum) which is a desertincluded species, and the Desert Slender Salamander (Batrachoseps aridus) which is a nondesert species in the desert-edge. 147

Unlike the reptiles, amphibians as a group in the North American Desert do not exhibit an inverse relationship between desertness (D) and riparianness (R). The flat-curve model for amphibians predicts a regression coefficient not significantly different from zero for riparianness (NR-RO) independent of degree of desertness (ND-DT). In the theory underlying the model there is an anamniotic versus amniotic hypothesis for the amphibian R on D pattern distinction from that of reptiles. It is at least clear from the test thus far that (with the exception of Batrachoseps and Hylactophryne) the desertarena amphibians in the North American Desert require reproductive surface water at sufficient depth .(> 1 cm) for a period > 10 days; they are the most clearly obligate riparian/wetland taxa in the desert herpetofauna and include both DI and DT desert species. Nondesert (ND) Species.-Eight amphibians and 28 reptiles, for a total 36 species, make the 25.2% nondesert taxa. A few are man's introductions, some are post-climax relicts. Some nondesert native species in the North American Desert established originally on riparian river and stream corridors of transported water well into the desert arena-inter-biome water from sources and sheds often from outside the desert region as well as transported water from within its overall geographic limits. Today that water may be permanent, seasonally intermittent, or torrentially ephemeral. In some cases the nondesert native species within now-arid environments were present where they are, as resident populations, under more mesic environmental conditions that preceded the more recent establishment of current desertscrub environments. Moreover, earlier less-arid climate in general is beyond reasonable doubt associated in various degrees with both mechanisms- riparian corridors and in-situ residency.

tional Park Service, National Science Foundation, The Arizona Foundation, and University of Arizona Research Fund 302787.

References
Axelrod, D. I. 1950. Evolution of desert vegetation in Western North America. In: D. I. Axelrod; Studies in late Tertiary paleobotany. Carnegie Inst. Wash. Publ. 90:1-23. Axelrod, D. I. 1958. Evolution of Geoflora. Bota. Rev., 24:433-509. the Madro-Tertiary

Axelrod, D. I. 1979. Age and origin of Sonoran Desert vegetation. Occ. Pap. Calif. Acad. Sci. No. 132:174. Dick-Peddie, R. B.; Hubbard, J. P. 1977. Classification of riparian vegetation. In: Johnson, R. R.; Jones, D. A., eds. Importance, preservation, and management of riparian habitats: a symposium. Gen. Tech. Rept. RM-43; 85-90. USDA Forest Service. Johnson, R. R.; Haight, L. T. 1987. Endangered habitats versus endangered species: a management challenge. Western Birds, 18:89-96. Little, E. L. 1950. Southwestern trees; a guide to the native species of New Mexico and Arizona. Agric. Handbook 9; 1-109. USDA Forest Service. Lowe, C. H. 1968. Fauna of desert environments. In: W. G. McGinnies and others, eds. Chapter VII: An inventory of geographical research on desert environments. University of Arizona, Tucson: Office of Arid Land Studies. Lowe, C. H.; Schwalbe, C. R.; Johnson, T. B. 1986. The venomous reptiles of Arizona. Phoenix, Arizona: Ariz. Game and Fish Dept. Reed, Jr., P. B. 1986. 1986 wetland plant list, Southwest Region. National Wetlands Inventory, WELUT-86/w13.07. U.S. Fish and Wildlife Serv., Washington. Shreve, F. 1917. A map of the vegetation of the United States. Geog. Rev., 3:119-125. Shreve, F. 1942. The desert vegetation of North America. Botan. Rev., 8:195-246. Shreve, F. 1951. Vegetation and flora of the Sonoran Desert. Vol. 1, Vegetation. Carnegie Inst. Wash. Publ. 591:1192. Stebbins, R. C. 1985. A field guide to western reptiles and amphibians. Boston: Houghton Mifflin Co.

Acknowledgements
Under methods are listed the names of many herptologists who have assisted me in the present work. I thank also R. Roy Johnson of the U.S. National Park Service who has been of able assistance in matters riparian and who is currently reporting on desert riparian systems for the U.S. Fish and Wildlife Service. The work has been supported by grants and contracts from the Na-

148

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

COYOTE CREEK (SAN DIEGO COUNTY) MANAGEMENT AND RESTORATION AT ANZA-BORREGO DESERT STATE PARK 1
David H. Van Cleve, Lyann A. Comrack and Harold A. Wier2

Abstract: Coyote Creek, along with its associated watershed in Anza-Borrego Desert State Park, is an extremely rich riparian system in the Colorado Desert of California. It provides habitat for the least Bell's vireo (Vireo bellii pusillus), is used as a critical summer watering site for the peninsular bighorn sheep (Ovis canadensis cremnobates), and was the site of major use and encampment by the Cahuilla Indians. Management activities have focused on maintaining or enhancing the integrity of the riparian system. Activities such as eliminating off-highway vehicle use and removing cattle with helicopters have been controversial, yet have demonstrated progress towards achieving the goals of protecting the riparian habitat, associated fauna, and cultural features of this Creek.

This region was also important to the Native American inhabitants. The Cahuilla Indians lived throughout and used the Coyote Creek area, as evidenced by the presence of 27 archeological sites in the Lower Willows area alone. These sites include 18 major aboriginal villages and smaller encampments. Also included are bedrock mortars, bedrock metates (milling slicks), and rock art sites. Undoubtedly, the availability of food, water, and shade made this area attractive to these early inhabitants. This area is also historically significant, since it was the route traversed by the Juan Bautista de Anza expedition in 1774-75 (Sampson, 1987). Much of the upland area surrounding the creek is still relatively inaccessible to humans. In fact, all but the existing routes of travel were classified as state wilderness in 1981 by the State Park and Recreation Commission. The "oasis" quality of the actual riverine system is extremely attractive to humans and animals. It is this same attractiveness which has led to its colonization and use by undesirable species.

Coyote Creek is the only reliably perennial creek in Anza-Borrego Desert State Park, San Diego County, California, and stretches 29 kilometers in length from the town of Anza to the edge of the Borrego sink near the town of Borrego Springs. Along its length are three distinct tracts of high quality riparian forest habitat known as Upper Willows, Middle Willows, and Lower Willows (the largest and most significant). These areas contain the following dominant plant species: narrow-leaf willow (Salix exigua), Fremont cottonwood (Populus fremontii), western sycamore (Platanus racemosa), arrowweed (Tessaria sericea), white alder (Alnus rhombifolia), mulefat (Baccharis glutinosa), honey mesquite (Prosopis glandulosa), and tamarisk (Tamarix ramosissima). Occasionally, fan palms (Washinglonia filifera) occur (Wier and Jones, 1986). One of the most significant residents of this area is the state threatened peninsular bighorn sheep. Approximately 150 bighorn rely on this drainage for food, water, breeding grounds, and cover. Sensitive bird species in Coyote Canyon include the state and federally endangered least Bell's vireo, black- crowned night-heron, green-backed heron, common yellowthroat, American kestrel, yellow-breasted chat, black-tailed gnatcatcher, blue grosbeak, downy woodpecker, willow flycatcher, yellow warbler, prairie falcon, red-shouldered hawk, and black-shouldered kite (Remsen, 1978; Tate, 1981)

Threats
Despite its status as part of a state park, its relative inaccessibility, and the wilderness designation of much of the surrounding landscape, the Coyote Creek area suffers from numerous threats to its ecological health and its value as a refuge to humans and sensitive fauna. These threats include the impacts of humans on the natural, cultural, and esthetic features (primarily through the effects of off-highway vehicles and archeological pothunters), and the effects of alien species on native flora and fauna. Off-Highway Vehicles Until March 1988, all vehicles were allowed to traverse the 29 kilometer route of Coyote Canyon from Borrego Springs to Anza, including the riverine routes. Although much of this jeep trail was located in upland areas, it passed directly through the three most significant riparian forests of Coyote Creek: Upper, Middle, and Lower Willows. Vehicular use in the last 40 years has changed

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Senior State Park Resource Ecologist and Associate State Park Resource Ecologist, respectively, California Department of Parks and Recreation, San Diego, California; and Senior Botanist, Michael Brandman Associates, San Diego, California.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

149

dramatically in this area. Originally used by jeeps and trucks as a route for sightseeing and nature enjoyment, Coyote Creek had, by the early 1980's, become the playground and speedway for motorcycles and All-Terrain Cycles (ATC). It was evident that the impacts of approximately 1000 vehicles on the riparian system during busy weekends were becoming intolerable. The Department of Parks and Recreation (Department) reevaluated its management practices towards off-highway vehicles because of the destruction of riparian habitat, suspected direct destruction of the nests of sensitive avifauna, deterioration of water quality, uncovering of archeological sites along with associated vandalism and theft, increased soil and bank erosion, the transformation of the area from a quiet sightseeing area to a noisy playground and other concerns (M. Jorgensen, 1987). Studies also showed that the peninsular bighorn sheep avoided the riverine area during periods of vehicular activity. During those periods of the year when air temperatures are above 35 degrees C, sheep require frequent access to watering sites. Their avoidance of the creek during peak vehicular activity caused stress in the sheep (P. Jorgensen, 1974). This study led directly to the implementation, in 1975, of an annual seasonal closure (June 15 September 15) of the entire Coyote Canyon watershed to all persons and vehicles. In July 1985, funding was secured to take the first step towards providing Lower Willows with permanent protection. Several alternatives were considered. These included extending the seasonal closure forward to March 15 to provide protection for the least Bell's vireo and other sensitive species during the breeding season, and implementing a permit process to limit the amount of traffic in the area. The alternative selected was a plan which proposed relocating the jeep trail out of the riparian forest into a xeric side canyon. Since this proposal provided resource protection and continued to allow vehicular access to the area, the Department felt that its project was fairly benign. The off-highway vehicle enthusiasts were incensed, however, and fought the proposal at every level possible. The proposal was heard by the State Park and Recreation Commission, which voted unanimously in favor of it and the concomitant proposal to classify the Lower Willows area as state wilderness. The San Diego Off-Road Coalition filed suit against the Department for not meeting the provisions of the California Environmental Quality Act (CEQA) for this project. After a lengthy process, the suit was adjudicated in favor of the Department in June, 1987. Construction of the bypass route was finalized in March, 1988, and opened to the public soon thereafter. This route now provides permanent protection to Lower Willows from the direct and indirect impacts of regular vehicular traffic. Horse and foot traffic is still allowed in

Lower Willows except during the annual seasonal closure. The other major action of the Department designed to lessen the impacts of motorized traffic on riparian resources of the area has been the issuance of a ban on all vehicles, except those which are street-legal, throughout the entire park. This ban, effective September 1, 1987, has effectively reduced the total number and type of vehicles accessing the Coyote Canyon area. Dirt bikes, all-terrain cycles, and many dune buggies may no longer be operated in Anza-Borrego Desert State Park. Although street-legal jeeps, trucks, and motorcycles still have impacts on the riparian forests of Middle and Upper Willows, these impacts have been greatly reduced. The types of vehicles and recreationists utilizing the area are now involved in slow-moving sightseeing instead of high-speed equipment testing. The physical and esthetic changes have become quite evident even in the short time since the institution of the ban; the traffic corridors have been filling in with thick stands of willow and tamarisk which provide additional avian habitat, and the scouring effect of rainstorms has eliminated many of the more obvious vehicle tracks. This regrowth was set back temporarily in May 1988 because of a wildfire in Lower Willows. This ban was also opposed vigorously by off-highway vehicle enthusiasts and groups in the proposal stage. Large letter-writing campaigns and strong political pressures were brought to bear upon the Department. It even became a major issue in the confirmation hearings of the new director of the Department. The proposal to ban off-highway vehicles was heard separately before the State Park and Recreation Commission, which again voted unanimously in its favor. At the time of this writing, the ban is in effect, and it is expected to remain so.

Threats to Least Bell's Vireo Various researchers have described the population status of the least Bell's vireo in California (U.S. Fish and Wildlife Service, 1986). Anza-Borrego Desert State Park provides extremely high quality desert breeding habitat for this species, with Lower Willows as an important population center. An intensive survey for least Bell's vireo was conducted in the park during 1986. A total of 32 territorial male vireos was found distributed in seven localities in the northern portion of the park. This figure is somewhat higher than estimates from past surveys, but is probably a reflection of more thorough and intensive field work rather than a population gain. Of this total, 9 were located in Lower Willows, representing 28 percent

150

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

of the total territorial males in the local desert population. Nesting success (percent of nests that produced vireo fledglings) was 54 percent, consistent with results found by researchers in coastal localities. Brown-headed cowbirds were prevalent in riparian habitat throughout the park. Cowbird nest parasitism was implicated as a significant factor reducing vireo productivity; at least 80 percent of the nests in Lower Willows had been parasitized (Wier and Jones, 1986). Future expansion of the least Bell's vireo into seemingly suitable habitat in Middle and Upper Willows will be dependent on the reduction of the cowbird parasitism threat. The Department began a vigorous cowbird reduction program in 1986 in an attempt to reduce significantly the local cowbird population. Six traps were built and set in Lower Willows, Riviera Farms (a private horse ranch), and the park's own Horse Camp. A total of 56 cowbirds was trapped, removed, and killed in the park in 1986. The dead cowbirds were frozen and donated to the San Diego Natural History Museum for use as study skins. The trapping results were somewhat disappointing, but can probably be attributed to a late start date that year (May 12th). The cowbird trapping program has continued through 1988 in Lower Willows. 170 cowbirds have been trapped and removed during the first part of this year's breeding season (Griffith and Griffith, 1988). The Department plans to continue the trapping program for several more years. The program will be improved as knowledge of cowbird dispersal patterns in Coyote Canyon increases. Tamarisk Eradication Tamarisk is a well documented problem species in the Southwest (Kerpez and Smith, 1987). As a phreatophyte, it evapotranspires tremendous amounts of water into the atmosphere. The result of this evapotranspiration is a reduction in the amount of water available for wildlife and for native riparian and waterhole plant species. Tamarisk tends to form dense, monotypic stands, outcompeting native vegetation. It is also less preferred as wildlife habitat than native riparian species. Tamarisk is pervasive in the Coyote Canyon drainage with several severe infestation areas. During May 1988, approximately 150 acres of riparian vegetation were burned in Lower Willows due to an escaped campfire. Two weeks after the burn, tamarisk had vigorously resprouted from underground rootstalks. Seedling establishment was also noted. The Department's response was to send workers into Lower Willows to cut and chemically treat the tamarisk. A park-wide tamarisk eradication program has been conducted since 1983, but had focused on remote springs and drainages where infestation has not been as severe. Past removal work included cutting the specimen with chainsaw or

loppers, and immediately applying the herbicidal agent Garlon-4 to the cut stump. Results have been encouraging. Following two successive seasons of removal work, a mortality rate of 95 percent on treated tamarisk has been achieved. Garlon-4, produced by Dow Chemical, is brushed or sprayed on the tamarisk stump immediately after the stem has been severed. Tamarisk resprouting has been minimal with the use of this chemical. Also, Garlon-4 is not a restricted pesticide, which makes it easier to obtain and use than other, more powerful, herbicides (Comrack, 1987). Techniques developed by park staff will be used on the Lower Willows tamarisk infestation problem in order to reduce the negative impacts of this weed on the riparian system. The fire consumed portions of riparian habitat used for nesting purposes by the least Bell's vireo. What direct effect the fire had on this population of vireos is unknown. Five territorial males have been noted in Lower Willows after the fire. We suspect that vireos may colonize nearby but previously unoccupied habitat at Middle or Upper Willows. Further surveys will be required to confirm this possible relocation. Another unfortunate side effect of the fire was its removal of the duff and litter layer in Lower Willows. Since this area had been used extensively by the Cahuilla Indians, the surface evidence of their occupation became very evident after the fire. Park rangers made an arrest for vandalism of a Cahuilla grave site, which included the possession of the remains of human cremation. Feral Cattle Livestock can cause many types of resource damage in riparian systems. They may compete with native wildlife species for food and water, destroy the riparian corridor through overgrazing, pollute surface water flows, compact soil, and increase erosion of streambanks. Additionally, domestic livestock carry contagious diseases which are deadly to the peninsular bighorn sheep (Jessup, no date). The presence of an increasingly large population of feral cattle in the backcountry of Anza-Borrego Desert State Park has long threatened the resource integrity of Coyote Canyon. Although grazing leases were canceled in 1971, remnant herds multiplied prolifically in the absence of any natural predators. Also, numbers of cattle increased from the illegal entry of branded cattle onto park property from neighboring ranches. Past efforts to rope, herd, and/or trap the cattle resulted in failure. An attempt was made in 1985 to introduce a bill in the State Legislature which would allow the Department to shoot feral cattle within the boundaries of the state park. This proposal met intense opposition from the California Department of Food and Agriculture and both the State and County Cattlemen's Associations.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

151

In the only avenue left unexplored, the Department initiated an aerial animal capture program, utilizing the professional services of Skydance Operations, Inc., of Minden, Nevada. Skydance provided a Hughes 500D helicopter especially equipped for aerial animal capture, pilot, net-gun operator, and other support staff in order to capture and transport the cattle successfully. Removal efforts were concentrated in Coyote Canyon due to the significance of its resource values and because past cattle damage to the drainage had been most severe. The field operation was broken into four phases as described below: Phase 1: A continuous supply of alfalfa was placed in key locations in Coyote Canyon a few weeks before the airlift operation began. Establishment of the feeding stations served to lure the animals into predetermined target range. Phase 2: Early morning reconnaissance flights served to locate large concentrations of cattle. During the actual animal capture, the pilot used the helicopter to herd the targeted bovine into the open. The animal was then captured with a net fired from a gun designed to be used from a helicopter hovering about 4 meters from the target. After firing, the 5 by 5 meter net settled onto and entangled the targeted animal. Captured cattle were hobbled with heavy leather straps, hoisted by hook to the helicopter, and flown out to holding pens erected in the park's Horse Camp. Phase 3: Ground crews subdued and untethered the cattle after they were unhooked from the helicopter. Veterinarians working independently from the Bighorn Research Institute in Palm Desert, California, collected blood samples and nasal swabs from each animal for diseases analysis. Cattle were then released to await transport. Phase 4: Through an arrangement with the Bureau of Livestock Identification, truckloads of cattle were taken from the Horse Camp to the town of Brawley. Animals were held for 14 days to determine ownership status. All unclaimed cattle were then sold at auction. The feral cattle removal operation began March 31, 1987 and continued through December 18 of that year. A total of 111 cattle were captured from Coyote Canyon and nearby Buck Ridge (M. Jorgensen, 1988). Although any animal capture operation poses some risk to the animals involved, all personnel and cattle escaped injury during transport. Disease analysis of blood and nasal discharge samples taken from the cattle was conducted independently by the Bighorn Research Institute. Preliminary results indicate that cattle have been exposed to many of the viruses which have decimated bighorn populations, including Para-Influenza III, Blue Tongue, Bovine Viral Diarrhea, Bovine Respiratory Synicital Disease, and Epizootic Hemorrhagic Disease (DeForge, 1987). It is hoped that this cattle removal pro-

gram will help ensure the long-term health of the Coyote Canyon bighorn population. In an effort to prevent further ingress of feral cattle into the park now that most existing animals had been removed, a fencing project was conducted in 1987. Approximately 13 miles of barbed-wire fencing were constructed at a cost of $194,000 on the western boundary of the park. Sites were selected which did not conflict with bighorn habitat and which were suspected to be major cattle ingress routes (M. Jorgensen, 1988).

Conclusions
The Department of Parks and Recreation has taken several important steps to identify and correct threats to the natural, cultural, and esthetic resources of Coyote Creek in Anza-Borrego Desert State Park. It is continuing to assess the populations of least Bell's vireo, peninsular bighorn sheep, and other sensitive species through contracts and volunteer programs. The Department is also continuing to monitor the success of the cowbird removal program through a contract and the success of the tamarisk removal and the Off-Highway Vehicle ban through studies conducted by park staff. It is vital that these monitoring efforts continue on a regular basis; they are expected to continue for several years. Even though these steps have been controversial at a variety of levels, they have been in concert with the legal mandates of the department and in conjunction with Departmental goals in the field of resource management. It is essential that established processes and laws be adhered to at each and every step. Too often, projects are stopped or delayed, not on their respective merits, but rather because the established process was not followed. In order to carry out resource management projects successfully, identification of and adherence to the process is an essential step of implementation.

Acknowledgements
We thank Anza-Borrego Desert State Park Naturalist Mark C. Jorgensen for contribution of field work, leadership, and coordination on all of these projects.

References
Comrack, Lyann A. 1987. Project Status Report on the 1986/87 Statewide Resource Management Program Project: Tamarisk Control, Anza-Borrego Desert State Park. Located at the Southern Region Office, Department of Parks and Recreation, San Diego, California.

152

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

DeForge, James. Director, Bighorn Research Institute. [Telephone conversation with Lyann A. Comrack]. 1987. Griffith, Jane, and John Griffith, Wildlife Biologists under contract to the Department of Parks and Recreation. [Telephone conversation with Lyann A. Comrack]. August, 1988. Jessup, David. No date. Diseases of Domestic Livestock which Threaten Bighorn Sheep Populations. California Department of Fish and Game, Wildlife Investigations Laboratory, Rancho Cordova, California. Jorgensen, Mark C. 1987. Off-Highway Vehicle Use in AnzaBorrego Desert State Park. California Department of Parks and Recreation Publication. Located at the Southern Region Office, Department of Parks and Recreation, San Diego, California. Jorgensen, Mark C., State Park Naturalist, Anza-Borrego Desert State Park. [Telephone conversation with David H. Van Cleve]. June, 1988. Jor ge nse n, Pa ul D. 1974. A S ur ve y of Ve hicle Use at a De se r t Bighorn Watering Area. Transactions of the Desert Bighorn Council. Kerpez, T., and N. Smith. 1987. Saltcedar control for wildlife habitat improvement in the Southwestern United

States. U.S. Department of the Interior. U.S. Fish and Wildlife Service. Resource Publication #169. Remsen, J. V. 1978. Bird Species of Special Concern in California. California Department of Fish and Game, Wildlife Management Branch. Administrative Report #78-1. Sampson, Michael P. 1987. An Archaeological Study of the Proposed Bypass Route, Lower Willows, Coyote Canyon, Anza-Borrego Desert State Park. Located at the Southern Region Office, Department of Parks and Recreation, San Diego, California. Tate, James, Jr. 1981. The Blue List for 1981. American Birds 35:1 (3-10). U.S. Fish and Wildlife Service. 1986. Least Bell's Vireo Management Plan. Portland, Oregon. Wier, Harold A., and Barry Jones. 1986. A Survey of the Birds of Riparian Habitats, Anza-Borrego Desert State Park, San Diego County, California, with Emphasis on the Least Bell's Vireo and Brown-Headed Cowbird. Westec, Services, Inc., Environmental Services Division. San Diego, California.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

153

SESSION D: RIPARIAN SYSTEMS AND FOREST MANAGEMENT


Forest streams create unique conditions along their margins, and even though a distinctive vegetation may not be present, managers have recognized the need to minimize disturbances in streamside corridors. Early management emphasis was on controlling sediment delivery and maintaining conditions that would support fisheries. The wider role that streamside vegetation plays in the ecosystem of forested basins and their streams is only now being fully appreciated. Many advances have occurred since the 1981 California Riparian Systems Conference. Two papers link forest management practices in and near the riparian zone with water temperatures and physical structure of the channel as it is affected by coarse woody debris. Another paper deals with the status of an unusual forest understory element, the Pacific yew. The final papers deal with management of riparian systems in watersheds where decomposed granitic soils and landslides exert significant influences. The papers in this session emphasize the high priority that riparian systems are starting to receive in forest management. Bruce McGurk Pacific Southwest Forest and Range Experiment Station, Forest Service Berkeley, California

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

155

PREDICTING STREAM TEMPERATURE AFTER RIPARIAN VEGETATION REMOVAL 1


Bruce J. McGurk2 Abstract: Removal of stream channel shading during timber harvest operations may raise the stream temperature and adversely affect desirable aquatic populations. Field work in California at one clearcut and one mature fir site demonstrated diurnal water temperature cycles and provided data to evaluate two stream temperature prediction techniques. Larger diurnal temperature fluctuations were observed in the water flowing through the clearcut than in the undisturbed area above the clearcut site. The mature fir forest also had a large diurnal water temperature variation. A 5.6C temperature rise was observed through a 380-m clearcut that exposed the stream channel, and Brown's equation predicted a change of 6.1C. A regression model underpredicted the maximum observed temperature by just under 2C at the clearcut site. A technique that includes the effect of shade recovery after timber harvest is suggested for use during long-range harvest planning. Early research determined that an important shading and sediment filtering role was played by the vegetation along channels, and this area was termed a buffer strip (Patton 1973). Management agencies have incorporated this concept by establishing special management areas along active stream channels that include the riparian zone and some amount of the adjoining hillslope. Limited. harvesting may be allowed in these streamside management zones (SMZ), which may vary in width depending on hillslope angle. Although equipment entry into the SMZ is discouraged, the restrictions do not prevent the removal of shade-providing vegetation from riparian zones. In addition, the Pacific Southwest Region (California) of the Forest Service, U.S. Department of Agriculture, has established Best Management Practices (BMP), which state that no adverse temperature impacts should occur to streams during harvests. The actual effectiveness of SMZ restrictions and other BMPs is not known due to the lack of detailed or long-term monitoring. Early efforts to predict stream temperature changes focused on predicting the maximum temperatures associated with peak summer conditions and low flows (Brown 1969). These early models were based on temperature changes caused by full exposure of the stream reach to the sun at the peak sun angle. By combining the site's latitude with field measurements such as stream temperature, channel width, depth, flow velocity, and an estimate of shading with estimates of potential cover reduction, likely temperature increases can be quantified. The estimated change in temperature, when added to the pre-harvest water temperature, provide an indication as to whether post-harvest temperatures might exceed the lethal limit for the resident fish. Other modeling approaches include empirical models that are calibrated for one geographic region, or detailed simulation models that require extensive data pertaining to the reaches to be modeled (Schloss 1985, USDA Forest Serv. 1984). The Schloss model is typical of a regression model and was developed in western Oregon to predict maximum summer temperature based on elevation, distance above the main channel, stream order, and shading. The USDA model was developed by the Forest Service to simulate stream temperature

Forest management can affect water quality and aquatic life, and riparian areas are both sensitive and easily disturbed. Streamside forest canopy removal allows direct sunlight to reach first- and second-order streams that were extensively shaded before timber harvest. Direct sunlight can increase stream temperature, which effects fish and aquatic insect species composition and growth (Feller 1981). Temperature also affects water quality parameters such as dissolved oxygen and the waste assimilation capacity of a stream. The effects of logging on stream temperature have been the subject of considerable research and numerous reviews (Brett 1956, Brown 1969, Patton 1973, Anderson and others 1976). Direct solar insolation was found to account for at least 90 percent of a stream's temperature change after clearcutting (Brown 1970). Salmon (Oncorynchus sp.), brown trout (Salmo trutta), and brook trout (Salvelinus fontinalis) prosper in streams that are between 10 and 18C, and if water temperatures exceed 24C they may die, depending on acclimation temperatures, pH, and dissolved oxygen (Patton 1973). The replacement of these high-value, cold-water fish species by warm-water fish has been associated with timber harvest.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Research Hydrologist, Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, Berkeley, California.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

157

response to multiple alternative harvest areas in a basin. It is a physical, energy budget-based algorithm, and has a time step that can range from 15-minute to hourly or daily intervals. Both direct and indirect (diffuse) shading is incorporated, as is stream aspect, topographic shading, groundwater influx and temperature, and flow into and out of the reach. The stream network is represented by sequentially estimating the outflow water temperature in each reach and using that information as the inflow temperature in the next downstream reach. A significant advantage to this model is its ability to handle partial shade, but obtaining the copious input data requires considerable field work. This paper reports on field work at two streams in California that evaluates Brown's stream temperature change prediction technique and an empirical equation developed in Oregon (Brown 1970, Schloss 1985). Both partial and complete riparian vegetation removal are analyzed. A modification of Beschta and Taylor's (1988) phased vegetation recovery system is proposed as part of a multireach accounting system for basins with multiple cutting areas.

T =

AN * .000167 Q

(1)

The coefficient contains the constants for the conversion of the flow, area, and insolation units to temperature. Because this model predicts a change in temperature, pre-project temperatures should be measured wherever harvests are planned. Streams should be visited during California's low flow and peak heat times of July, August, and September. A simple pocket thermometer could yield representative data for several small basins with a moderate amount of effort, using measurements taken between noon and 1500 hours. Peak temperatures occur due to the interaction of declining streamflow and insolation, in spite of the decline of insolation after June 21. The parameter A reflects the new channel area that will be exposed due to forest harvest, but topography, channel aspect, and harvest design also have a role in determining A, so subjective judgments may be needed. If 35 percent of the cover in a 100 m zone along the southside of a channel is to be removed, it may be reasonable to equate this to complete removal from about 30 m of channel. Table 1 Average values of net solar radiation
absorbed by water surfaces in middle latitudes for a range of exposure times (cal/cm2-min) (after List 1951, Brown 1974). Water Travel Latitude (degrees) Time (hours) 35 40 45 2 1.30 1.28 1.22 4 1.25 1.22 1.17 6 1.19 1.14 1.11 8 1.09 1.06 1.00

Temperature Prediction

Model Selection Model selection should be based on the size of the area of concern and on the intended use of the water temperature prediction. Because the typical forestry use is to assess the effect of timber harvest, grazing, recreation, or road construction on large land areas, the complex and data-hungry physical simulation models are inappropriate. Empirical (regression) models may be appropriate if one has been developed for the local area of interest. In most cases, however, a relatively simple model based on the physical processes relating stream surface exposure to sunlight is most appropriate.

Exposed Surface Models Exposed surface models combine a few crucial types of field data with tabular data dependent on site location (Brown 1969). This type of model uses only physical constants and field measurements, so it is not an empirical, "calibrated" model. Changes in water temperature T(C) increase directly in relation to new stream surface area A (m2) that is exposed and insolation N (cal/cm2min), and inversely with streamflow Q (m3/s):

Solar loading N is dependent on season, latitude, and the length of time that the water is in an exposed area. California's National Forests range from 34 to 42 latitude, so N values for the appropriate latitudes have been estimated (table 1). N values could be reduced by about 1 percent for each week after July 1 to account for the seasonal decrease in insolation, but such minor adjustments are probably not warranted due to the inherent errors in area and discharge estimates. The travel times for the 160 m to 400 m openings typical of National Forest System operations and stream gradients are between 1 and 2 hours, so the N values for 2 hour travel times in table 1 should be used for most small streams. The final requirement for equation 1 is discharge volume, and small mountain streams are difficult to gauge

158

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

accurately due to shallow depths, turbulence, and sidepool areas. If a small current meter is available, measurement of cross-sectional areas and water velocities can provide reasonably accurate results. Alternately, dye or floating objects such as oranges can be used but accuracy will suffer. If objects such as sticks are used, the velocity should be multiplied by 0.8 to correct for the vertical velocity profile of the stream. Cross-sections should be selected to minimize stagnant water pools near the stream's edge or discharge can be overestimated by 50 to 100 percent. Empirical Prediction Empirical equations can be developed by regressing stream temperature on basin, cover, and stream characteristics (Schloss 1985): T = 11.9 0.0013E + 0.206L + 0.676R + 1.814(S/50 + 1) where: T = maximum summer stream temperature (C) E = midbasin elevation (m) L = distance from junction of next higher-order stream (km) R = stream order S = shade percentage (percent) Standard deviation = 1.7C. Equation 2 was calibrated for forested basins in western Oregon that were below 610 m elevation. Unlike equation 1, this technique predicts maximum temperature rather than temperature change. The stream order and channel distance factors are measured on US Geological Survey 7.5 quadrangle maps. The channel length is the distance from the area of interest to that stream's juncture with the "main" channel. The shade code is the percentage of channel that has less than "complete" shade within 1600 m upstream from the point of interest. (2)

the slash disposal burn got out of control and destroyed most of the remaining near-stream vegetation. These actions produced a 380-m section of stream that had almost no shading. Field instrumentation consisted of water temperature, air temperature insolation, humidity, and wind instruments. Ten water temperature probes were placed in the unshaded channel, one probe was 70 m upstream of the cut, and probes were placed 35 m and 90 m downstream of the clearcut area. Except for a hygrothermograph and rainfall collector, all readings were collected electronically at 15-minute intervals. The site was monitored for 48 hours between August 31 and September 2, 1983. Approximately 1.3 cm of rain fell during the afternoon and evening of August 31, but September 1 and 2 were warm with clear skies. Peak air temperatures were 29C on September 1 and 32C on September 2. The average discharge during the study interval was 18 l/s (0.6 ft3/s).

Site Descriptions and Field Methods

McGill Creek A clearcut site was identified 3 km north of Iron Canyon Reservoir along McGill Creek at an elevation of 915 m (figure 1). Iron Canyon Reservoir is in the Shasta National Forest and is 61 km northeast of Redding, California. McGill Creek is a south-draining secondorder stream, with a slope of 3.5 percent, that passes through an 8-ha clearcut. The timber operator removed nearly all of the timber on both sides of the stream, and

Figure 1California map pinpointing McGill Creek clearcut site and Teakettle Creek mature fir site where field tests took place.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

159

Table 2 - Average water temperatures and meteorological data for McGill Creek near Redding, and for Teakettle Creek near Fresno, California.
Date Time Entry Water Exit Water Temp.(C) Temp.(C) 12.0 11.8 11.6 11.5 11.7 12.2 12.1 11.6 11.0 10.8 11.3 12.3 12.0 10.7 11.2 11.6 13.9 16.7 15.6 12.2 10.9 10.8 14.1 17.4 Temp Diff (C) 0.0 -1.1 -0.4 0.1 2.2 4.5 3.5 0.6 -0.1 0.0 2.8 5.1 Air Temp. (C) 12.8 10.3 9.4 10.0 15.9 20.6 16.9 7.0 5.2 8.2 22.9 27.0 Insolation (C) 0.04 .0 -.01 .06 .65 .82 .21 -.05 .0 .02 .82 .89 Windspeed (cal/cm2-min) 0.6 .3 .1 .1 .5 .7 .7 .4 .6 .6 .6 .8

McGill Creek 8/31 15-18 18-24 9/01 0-6 6-9 9-12 12-15 15-18 18-24 9/02 0-6 6-9 9-12 12-15 Teakettle Creek 8/26 15-18 18-24 8/27 0-6 6-9 9-12 12-15 15-18 18-24 8/28 0-6 6-9 9-12 12-15

11.1 9.0 7.5 7.1 9.5 12.0 11.3 9.2 7.8 7.5 9.7 12.0

10.6 8.9 7.4 7.0 9.2 11.3 10.8 9.0 7.6 7.4 9.4 11.3

-0.5 -0.1 -0.1 -0.1 -0.3 -0.7 -0.5 -0.2 -0.2 -0.1 -0.3 -0.7

13.8 6.8 5.1 9.3 24.4 22.3 14.6 7.4 5.8 9.1 23.4 21.2

0.01 .0 .0 .02 .69 .38 .0 .0 .0 .01 .70 .32

0.5 .8 .8 .8 .4 .5 .4 .0 .0 .0 .5 .3

Teakettle Creek The Teakettle site is on the Sierra National Forest at an elevation of 2100 m. It is in the Teakettle Experimental Forest, on the southeast flank of Patterson Mountain and 66 km east of Fresno, California. Teakettle Creek is a southeast-facing, second-order drainage with a slope of 8 percent that passes through senescent red fir. Although some clearings exist due to the presence of 10 m by 40 m wet meadows, the combination of extensive shrub growth and the 50- to 80-m fir trees exclude most direct exposure from sunlight. A shading survey produced an estimate of 80 percent canopy cover. The field instrumentation at Teakettle was similar to that used at McGill Creek. Approximately 380 m of stream channel was monitored with 11 water temperature probes, and the other instruments were sited along the stream channel. Peak air temperatures were 27C on August 27 and 25C on August 28, 1983. The average discharge during the study was 39 1/s (1.3 ft3/s).

Measurement Accuracy All thermistor probes were calibrated by measuring their resistances in three baths of known temperature that spanned the expected measurement interval. The agitated water baths were measured using a precision thermometer accurate to 0.1C. A separate polynomial equation was developed for each probe. Replicate stream temperatures were measured by placing two probes within 2 cm of each other at a single random spot at both McGill and Teakettle Creeks. The mean difference around the replicates and the confidence limits around the difference between any two probes were as follows: Mean difference (C) McGill Teakettle 95 Pct. Confidence interval (C) 0.16 0.3 0.20 0.4

Based on these confidence intervals, observed water temperature values that differ by less than 0.8C must be considered to be the same.

160

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

much higher than at McGill. Teakettle Creek's water temperature is lower than the "natural" case at McGill, a situation that may be due in part to the 1100 m elevation difference. The higher variability at Teakettle may be due to the upstream springs that supply the stream. At McGill, side seeps were common along the channel, and their presence would add to the diurnal variation due the shallow, marshy flow that was exposed to the sun. Average insolation, windspeed, and air and water temperatures illustrate some of the differences between the sites (table 2). The difference between the 3hour average water temperature entering and leaving the clearcut was about 5C at McGill, but the stream actually lost heat in the measured reach at Teakettle. The open site had both larger daytime energy inputs and larger nighttime energy losses due to the lack of a canopy. The total allwave flux was 790 cal/cm2 at McGill and 383 cal/cm2 at Teakettle. Windspeeds at the two sites were roughly equivalent, but little wind movement would have been possible in the natural channel areas above the clearcut at McGill due to the dense vegetation close to the water surface. The canopy at Teakettle, however, is much higher, allowing typical diurnal wind patterns.

Exposed Surface Water Temperature Prediction The McGill Creek site was well suited to Brown's (1970) model for predicting temperature. Channel area was calculated using an average width estimated by measurements at six locations along the channel. In addition to the 1.9-m width along the 380 m of channel, there were also eight small pools that had been constructed for gradient control and to allow sediment to settle. The pools added 111 m 2 to the 483 m 2 of channel surface area, so the estimate of the total exposed area was 594 m2. McGill Creek is at 41 latitude and the water travel time was about 1 hour (velocity = 0.1 m/s, so the N factor (equation 1, table 1 for 2 hours) equals 1.27. The average discharge, as measured by both current meter and dye velocity/crosssection measurements was 0.019 m3/s. The calculated temperature change was 6.6C, and the observed water temperature increase through the cut area was 5.0C on September 1 and 5.4C on September 2. Both the calculated and observed temperatures are estimates that include measurement errors. For equation 1, the area term may have about a 25 percent error, the insolation error may be 20 percent, and the discharge error may be 50 percent. The combined effect of these errors suggests that the predicted value of 6.6 C is the "best guess" in a range of predicted temperature increase that extends from 3C to 20C. Some decrease in the error band may be obtainable with extreme diligence

Figure 2Recorded diurnal water temperature at the two field test sites, August 31 September 2, 1983.

Results and Discussion


The diurnal water temperature at the two sites share a similar pattern, but there are important differences (figure 2). The air temperatures at both sites peaked at between 24C and 28C. The McGill Creek water temperature was 17C on September 1 and 18C on September 2, but Teakettle water temperatures peaked at 12C on both August 27 and 28. Although both sites produced a sine-shaped temperature pattern, the amplitude varied at the two sites. At McGill Creek, one can hypothesize that the diurnal variation of temperature would be very small in the natural system. Sensors 65 m above and at the upper margin of the clearcut show very small diurnal variations (figure 2). This small variation is due at least in part to the dense shade provided by the willow and alder that choked the channel upstream of the clearcut area. The mature fir forest at Teakettle Creek provided the channel with only 80 percent cover, and the overstory was

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

161

during data collection. The errors associated with the probe measurements are discussed above. Part of the difference between the predicted and observed values could be due to the decreased solar strength in early September as compared to the peak strength associated with the June 21 summer solstice. Peak insolation at the Central Sierra Snow Laboratory, near Soda Springs, California, declined by 8 percent during that interval. Assuming that the same pattern is followed at McGill Creek, the decreased solar input accounts for 0.5C, dropping the predicted change for the actual measurement period to 6.1C. The remaining difference could be due to the shrubby vegetation along the channel, ground water inflow, or errors in the stream area or discharge measurements. Equation 1 is sensitive to errors in discharge estimation, especially on small streams with low total flows. If the 0.019 m3/s value is varied by 10 percent, the initial predicted temperature change (6.1C) changes to 5.6C or 6.8C. Typical current meters are accurate to approximately 5 percent (USDI Bur. Reclam. 1975), and errors as large as 50 percent are likely in small channels due to lateral turbulence and shallow depths. The largest 3-hour average insolation values in table 2 for McGill Creek are 35 percent less than a solar loading value of 1.26 estimated for a site at 41 latitude from table 1. The instantaneous net allwave values measured at McGill Creek peaked at 1.1 cal/cm2-min. If the tabular value is reduced by the 8 percent seasonal factor, the value becomes 1.16 cal/cm2-min, a value that is only 0.06 cal/cm2-min different than the observed value. The Teakettle Creek site is not as well suited for the application of Brown's technique as was McGill Creek. Although no new channel area had been exposed due to harvesting, the 80 percent canopy cover implies that 20 percent of the stream is exposed to insolation. The channel survey yielded an average width estimate of 3.3 m and a length of 380 m, so there is 1254 m2 of surface area and 251 m2 of the total is exposed. The insolation value for a 2-hour travel time at 37 latitude, corrected by the 8 percent seasonal factor, is 1.19 cal/cm2-min. The observed discharge was 0.037 m3/s, so the predicted temperature increase was 1.4C. The field results show a top-to-bottom temperature decrease of almost 1C on both days. Due to the measurement and prediction error factors mentioned above, there is no difference between estimated and observed values, but the divergence is interesting. The decreasing water temperature is counterintuitive in that no large open areas above the measurement site were present from which the stream was recovering. Further, the water temperature at the top of the reach was already rather low for the peak summer heat period.

The diurnal variations at the two sites were markedly different. The Teakettle site had diurnal variations of 4C, but the undisturbed portion of the McGill site had diurnal variations of 1.2C. This difference may be due to the lack of low shrub cover at Teakettle versus very dense willow and alder at McGill Creek. The water temperature at Teakettle declined markedly during the night, and this pattern was not seen at McGill in spite of similar air temperatures. Empirical Temperature Prediction McGill Creek's elevation is 915 m (E), it is a second order stream (R), and the site is 2.4 km (L) from Iron Canyon Reservoir. The 380 m of clearcut area produces an S value of 24 percent because the remainder of the channel was shaded. It is likely that there would be less overall effect if the clearcut area was split into two portions at either end of the 1600 m effective distance, but this method lumps all partial or unshaded areas into a single ratio. Equation 2 predicts a summer maximum temperature of 15.3 C. Compared with observed maxima of 17C and 18C, the predicted values are surprisingly close. As a second test at McGill Creek, a prediction can be made for the undisturbed area above the clearcut. The shading factor becomes zero and the channel length changes to 2.5 km. The predicted maximum water temperature is 14.4C , and the observed maximum was less than 12C. Teakettle Creek is at 2100 m elevation, is a firstorder stream, and the site is 3 km from the Kings River. Using a shade factor of 20 percent, the predicted summer maximum was 13C with a standard deviation of 1.7C. The observed maximum water temperature was 12.3C, not significantly different than the predicted value. Because Teakettle is further from Oregon and higher than McGill, plus has no real clearcut areas, the correspondence between the observed and predicted temperatures is surprising. Although these three cases are not an adequate evaluation of Schloss' equation, they do show both the promise and the danger associated with an empirical approach. An equation that was calibrated for a geographic area could be very useful and reasonably accurate. Indiscriminant use, however, could conceal problem situations that deserve closer attention. Heat Loss Elevated water temperature may decrease once the heat input disappears. At McGill Creek, a sensor was located 130 m below the clearing. After the Creek flowed under the dense canopy cover for this distance, the peak

162

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

temperature listed above and shown in figure 2 decreased by 1 or 1.5C. Heat was lost to the streambed or to the air, but it is not known if this rate of heat loss continued or if the water returned to its original temperature at some downstream point. Many streams lose heat and return to their elevation-, flow-, and groundwaterinfluenced base temperature within 1.6 km of their exit from a disturbed area (Schloss 1985).

Table 3 - Shade recovery calendar for aiding the scheduling of


timber harvests within a basin (after Beschta and Taylor 1988). Harvest Year1 1960 65 70 75 80 85 90 95 2000 05 10
1

Harvest Event A 1 1 .6 .3 0 B 1 1 .6 .3 0 C 1 1 .6 .3 0 D 1 1 .6 .3 0 E 1 .6 .3 0 F 1 1 .6 .3 0 G 1 1 .6 .3 0 H 1 1 .6 .3 0 I 1 1 .6 .3 0 Total 2.0 3.6 2.9 2.5 2.6 3.6 3.9 2.5 1.2 0.3 -

Multiple Harvest Areas Although one harvest may have only a small effect on stream temperature, multiple harvests within a few years might produce a "cumulative effect" on downstream temperature. For an Oregon watershed following clearcut harvesting, little shade recovery occurred within 5 years after stream banks were cut, but a linear and total recovery occurred during the subsequent 15 years (Beschta and Taylor 1988). Although some stream temperature models have multicut, multiyear capability, the data requirements preclude their use on basins with miles of channels and numerous subbasins (USDA For. Serv. 1984). A tabular recovery analysis for basins could aid the harvest planning by explicitly incorporating shade recovery information (table 3). The table incorporates a 20-year vegetation growth cycle, and the procedure uses an index that varies from 1 (full effect) to 0 (no effect) to represent the loss of shading due to harvest if any canopy cover is removed from the riparian zone. After 20 years, the index returns to zero as stream shading recovers. In table 3, harvest E occurred near 1960, A occurred near 1965, and B and D occurred near 1970. The column labeled "Total" is the sum of the horizontal coefficients, but the value that should be considered to be a cumulative effect threshold is unknown. If the average riparian timber removal is 50 percent along the associated 300 m of channel and five harvests occurred within a 5-year period, a value of five in the "Total" column might represent 750 m of clearcut stream channel. The incorporation of this technique during the harvest plan could provide a feedback system such that predicted increases in estimated stream temperatures would increasingly restrict the removal of shading vegetation. A monitoring plan that proceeded concurrently with the harvest would provide valuable information on temperature effects.

Assign harvests to nearest 5-year date.

Conclusions
The exposed surface area model (Brown 1970) for predicting stream temperature may be a good choice for land managers because it requires a minimum of field data that are relatively simple to obtain. If a sufficient data base exists within a region or can be collected over time, an empirical model will simplify maximum temperature prediction associated with shade removal. Field data from both a clearcut and a mature fir site were used. A predicted temperature change of 6.1C compared well with an observed change of 5.4C at a 380 m clearcut site. The prediction equation is sensitive to streamflow, a factor that is known to be difficult to measure with less than at least 5 percent error. The 80 percent-shaded Teakettle site yielded a predicted increase of 1.4C compared to an observed decrease of almost 1C. Results from the empirical model were 2C lower than the observed water temperatures in the clearcut portion of McGill Creek and 2C higher than the undisturbed area (Schloss 1985). The regression model's prediction nearly matched the fir site's water temperature of 12C. If data were collected for several areas of California and used to calibrate a model with similar structure, greater consistency might be achieved. This type of model has the advantage of requiring no additional field data once the coefficients are estimated. A shade recovery accounting system was proposed for use during the National Forest System harvest planning process. The system assumes channel cover is regained in 20 years and offers the planner a way to avoid overscheduling harvests in a basin and producing an adverse cumulative temperature effect.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

163

Acknowledgments
I thank Keith Maclntyre and Michael Pack for their assistance in collecting the field data. The staff of the Supervisor's office, Shasta-Trinity National Forest, provided maps and other much-appreciated assistance.

Brown, G.W. 1974. Forestry and water quality. Corvallis: Oregon State University Book Stores; 74 p. Feller, M.C. 1981. Effects of clearcutting and slashburning on stream temperature in southwestern British Columbia. Water Resources Research 17(5):863-867. List, R.J. 1951. Smithsonian Meteorological Tables, 6th ed. Washington: Smithsonian Institution Pub. 4014, Vol 114. 527 p. Patton, D.R. 1973. A literature review of timber harvesting effects on stream temperature. Research Note RM-249. Ft. Collins, Colorado; Rocky Mountain Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 4 p. Schloss, A.J. 1985. A predictive model for estimating maximum summer stream temperatures in western Oregon. Technical Note 370, Denver, CO: Bureau of Land Management Service Center, U.S. Department of Interior; BLM/4A/PT/85-006-4341; 8 p. U.S. Department of Agriculture, Forest Service. 1984. Temp84: A computer model for predicting stream temperatures resulting from the management of streamside vegetation. WSDG-AD-00009, Ft. Collins, CO; Watershed Systems Development Group, U.S. Department of Agriculture, Forest Service; 79 p. U.S. Department of the Interior, Bureau of Reclamation. 1975. Water Measurement Manual. 2nd ed., Denver, CO; U.S. Department of the Interior, Bureau of Reclamation; 327 p.

References
Anderson, H .W.; Hoover, M.D.; Reinhart, K.G. 1976. Forests and water: effects of forest management on floods, sedimentation, and water supply. Gen. Tech. Report PSW18. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 155 p. Beschta, R.L.; Taylor, R.L. 1988. Stream temperature increase and land use in a forested Oregon watershed. Water Resources Bulletin 24(1):19-25. Brett, J.R. 1956. Some principles in the thermal requirements of fishes. Quarterly Review of Biology 31:75-81. Brown, G.W. 1969. Predicting temperatures of small streams. Water Resources Research 5(1):68-75. Brown, G.W. 1970. Predicting the effect of clearcutting on stream temperature. Journal of Soil and Water Conservation 25(1):11-13.

164

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

COARSE WOODY DEBRIS ECOLOGY IN A SECOND-GROWTH SEQUOIA SEMPERVIRENS FOREST STREAM1


Matthew D. O'Connor and Robert R. Ziemer2 Abstract: Coarse woody debris (CWD) contributes to high quality habitat for anadromous fish. CWD volume, species, and input mechanisms was inventoried in North Fork Caspar Creek to assess rates of accumulation and dominant sources of CWD in a 100-year-old second-growth redwood (Sequoia sempervirens) forest. CWD accumulation in the active stream channel and in pools was studied to identify linkages between the forest and fish habitat. CWD accumulates more slowly in the active stream channel than on the surrounding forest floor. Of CWD in the active channel, 59 percent is associated with pools, and 26 percent is in debris jams. CWD associated with pools had greater mean length, diameter, and volume than CWD not associated with pools. The majority of CWD is Douglas-fir (Pseudotsuga menziesii) and grand fir (Abies grandis). CWD entered the stream primarily through bank erosion and windthrow. The estimated rate of accumulation of CWD in and near the stream was 5.3 m3. Selective addition of CWD to stream channels to compensate for reduced inputs following timber harvest could maintain or enhance fish habitat.

Study Area
The 508-ha North Fork Caspar Creek (Caspar Creek) watershed, in the Jackson Demonstration State Forest, Mendocino County, California (fig. 1), was clearcut and burned 90 to 100 years ago. A splash dam was constructed in the upper one-third of the watershed, and was periodically breached to transport cut logs. Native runs of steelhead trout (Salmo gardnerii gardnerii) and coho salmon (Oncorhynchus kisutch) utilize the full length of Caspar Creek below the splash dam site. The Caspar Creek watershed is underlain by Franciscan graywacke sandstone. Slopes are steep and mantled with deep soils in which large rotational landslides are common. The stream is in an inner gorge 10 to 20 m deep. Bedrock is exposed locally. Alluvial terraces, 1 to 2 m above the channel, form the 15- to 25-m-wide valley floor. The channel gradient averages 2.0 percent in the study reach.

Coarse woody debris (CWD) greater than 10 cm diameter in streams contributes to high quality habitat for anadromous fish (Bisson and others 1987). Management options for anadromous fish in north-coastal California could be expanded through analyses of CWD ecology. The abundance of CWD in streams flowing through oldgrowth redwood (S. sempervirens) has been documented (Keller and others in press). Input processes, accumulation rates, and dynamics of CWD have been studied in the Pacific Northwest (Lienkaemper and Swanson 1987). This paper reports a study that estimated CWD abundance and its relative proportions in the riparian zone and in the channel of a stream following through second-growth S. sempervirens. Data for CWD input mechanisms, tree species, and numbers of CWD pieces were also collected.

Figure 1 Caspar Creek Experimental Watershed

1 2

Presented at the California Riparian Systems Conference, September 22-24, 1988; Davis, California. Hydrologist; and Principal Research Hydrologist, Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture, Arcata, Calif.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

165

At the lower end of the study reach, the mean annual flood occurs between November and May at about 0.5 m flow depth, with a discharge of about 3.1 m3 flow depth of 1 m (discharge 8.5 m3/sec) would be expected once every 25 years at this location in the watershed. During late summer, the minimum streamflow averages about 0.001 m3/s.

CWD volume was calculated as for a cylinder. Ratio estimators (Cochran 1977) were used to estimate total CWD volume in each zone.

Methods
CWD Volume

The volume of CWD was estimated in three zones defined as follows: Classification Effective zone Description Within the active channel, up to 0.5 m above the water surface or the channel bed. Suspended 0.5 m or more above the channel, extending laterally 1 m beyond the vertical channel bank. Stream channel and terraces; lateral boundary is steep slope

Potential zone

Valley floor

These classifications are hierarchical. The effective zone lies below, is smaller than, and exclusive of the potential zone; the valley floor includes both the effective and potential zones (fig. 2). Our definitions of effective and potential CWD are similar to those used by Swanson and others (1984) and partition CWD into present and future supplies. Functionally, effective zone CWD was that in contact with the water column during the mean annual flood. Potential zone CWD was that transferable to the effective zone by breakage or bank erosion. CWD volume in the potential and effective zones was estimated by measuring lengths and end diameters of CWD pieces with a stadia rod and log calipers in random sample plots. The 1850-m study reach was divided into 185 plots 10 m long and three 600-m sub-reaches (figs. 1 and 2a). To ensure even distribution of sampling, 20 plots were selected from each sub-reach. Portions of CWD pieces outside the plot were ignored. Plot size varied with channel width; the area within each plot was estimated as the product of length (10 m) and the mean of 3 active channel width measurements.

Figure 2 Plan and cross-section views of sampling zones in Caspar Creek, California. (a) Plan View: Relationship of geomorphic features to sample plot and linetransect. Lateral plot boundary parallels streambank at a distance of 1 m. (b) Effective zone width determined by active channel. Effective zone contained in potential zone; potential zone contained in valley floor.

166

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

The line transect technique (Van Wagner 1968) was used to estimate quantities of CWD on the valley floor. Random transects extended 50 m diagonally across the valley floor in a "zig-zag" pattern (fig. 2a,b). CWD volume per unit area of valley floor was estimated as:

railroad ties. Wind fragmentation included pieces broken from a snag. Landslide CWD was defined as tree falls with rootwads in landslide deposits. The unknown category was used when no input mechanism could be identified.

V = (( )2 /8L * (d)

Results
CWD Volume and Density We estimate that 1180 m3 of CWD lies on the valley floor in the 1850-m study reach (table 1), 37 percent (437 m3) of which is in either the potential zone (25 percent) or the effective zone (12 percent). Seven percent of CWD in the study area was pool-associated.
Table 1- Coarse woody debris volume and density at Caspar Creek, California.

where L was the transect length (50 m) and d was diameter of CWD in meters. Base-flow fish habitat was inventoried by Decker and Lisle (work ongoing, unpublished data for June 1987) using the habitat classification scheme of Bisson and others (1982) modified for use in northern California. Pools were classified according to the cause of formation, e.g., scour around a rootwad, deflection by bedrock outcrop. CWD that affected inventoried pools ("associated" became a subset of effective zone CWD. CWD that created a scour zone that aided pool formation, or was in a pool, or appeared to be in the water column above a pool during the average annual flood was classified as "pool-associated." CWD in the effective zone was also classified according to whether it was part of a woody debris jam, defined as aggregations of at least three CWD pieces. Whether CWD pieces appeared to have been mobilized by streamflows was also noted.

Valley Floor Potential + Effective Potential Effective Pool-Associated Debris Jams Effective Pool-Associated

Total +S.E. (m3) 1180 182 437 41 294 34 143 16 84 14 57 20 38 8

Area (ha) 3.03 1.29 1.29 0.82

Density (m3/ha) 389 339 228 174

CWD Species
Tree species from which CWD originated was recorded. Because of the difficulty of distinguishing Abies grandis from Pseudotsuga menziesii, these were lumped together as "fir." S. sempervirens was the only other species found in substantial quantity. A third category of "other" species included known species with minor representation and CWD for which the species could not be determined. CWD accumulates in the combined potential and effective zones at the same rate that it accumulates on the valley floor. CWD density on the valley floor was 389 m3/ha, and in the combined potential and effective zones, 339 m3/ha (table 1). The 95 percent confidence interval of the difference includes 0 (50 + 133 m3/ha). CWD accumulates in the effective zone at a slower rate than on the valley floor. CWD density in the effective zone was 174 m3/ha. The 95 percent confidence interval of the difference in density between the effective zone and the valley floor did not include 0 (215 + 124 m3/ha). In the effective zone, 59 percent of the CWD volume was pool-associated.

CWD Sources Six sources of CWD were identified: (1) bank erosion, (2) windthrow, (3) logging debris, (4) wind fragmentation, (5) landslide, or (6) unknown. Bank erosion included rootwads or trees with attached rootwads which collapsed from the banks into the channel. Windthrow was defined as tree falls with attached rootwads originating from the valley floor and adjacent slopes. (Tree falls attributed to bank erosion, simply because of the trees' proximity to the stream, might have been caused by wind.) Logging debris included pieces with cut ends, burn scars, or milled edges, such as bridge parts and

CWD Species Differences in species composition were statistically significant (95 percent confidence level) for the combined data from the potential and effective zones, where fir was more abundant than redwood, and redwood was more abundant than other CWD. Despite limited statistical evidence, we are confident that fir-derived CWD is

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

167

about two times more abundant than redwoodboth volume and number of piecesand that other species are insignificant for practical purposes (fig. 3). We attribute the lack of

statistical evidence to the estimators used, the high frequency of null data in sample plots, and the inability to identify species of rotted pieces.

Figure 3 - Coarse woody debris species as percentage of sample volume (a) and pieces (b) at Caspar Creek, California. 168

Figure 4 - Coarse woody debris as a percentage of sample volume (a) and pieces (b) at Caspar Creek, California.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

CWD Sources The high proportion of CWD volume and number of pieces classified as "unknown" (fig. 4) were due to the long time since some of the trees fell, and the tendency of fallen trees to be fragmented and transported by the stream. Because of this tendency, differences between sources were not statistically significant. Despite these classification problems, most CWD volume in the potential zone (60 percent) was windthrown fir. Many fallen trees uprooted from the lower slopes of the inner gorge criss-crossed the valley floor and were suspended above the stream channel. The high percentage of CWD from logging found on the valley floor (36 percent of volume), came from a few large diameter redwood pieces. Only 8 percent of the pieces on the valley floor were logging remnants. In the effective zone and its pool-associated component, windthrow and logging contributed approximately equal proportions of CWD. Bank erosion contributed about as many pieces of CWD as windthrow and logging combined. Debris jams, which develop after CWD input, are of equal or greater importance than direct input processes (windthrow and bank erosion) as pool-associated debris. Debris jams accounted for 27 percent of the volume and 56 percent of the pieces of pool-associated CWD. Diameter and length affected the likelihood of CWD being associated with pools and being transported by the stream. Mean length, mean diameter, and mean volume of CWD was significantly greater, at the 95 percent level of confidence, when associated with pools than when it was not. The mean volume of CWD associated with pools was 0.17 m3 compared with 0.07 m3 for CWD not associated with pools. Mobile CWD had significantly smaller mean values of length, diameter, and volume than non-mobile CWD. The mean volume of mobile CWD was 0.06 m3, compared with 0.16 m3 for non-mobile CWD.

m3/ha to 4500 m3/ha. The CWD density we measured in Caspar Creek is comparable to that for P. menziesii old-growth forest streams in the Klamath Mountains mean of 282 m3/ha and range from 10 m3/ha to 1200 m3/ha for 11 sample reaches (Harmon and others 1986).

CWD Ecology The distribution of CWD can be visualized as follows: of each 10 m3 of CWD within the limits of the valley floor, about 2.5 m3 are in the potential zone, 1.2 m3 are in the effective zone, and 6.3 m3 are lying on that portion of the valley floor outside of these two zones. Within the effective zone, 0.7 m3 of CWD are associated with pools. We attribute this distribution in large part to the dimensions of the stream channel and the mechanics of tree fall. Because the channel lies within vertical streambanks 1 to 2 m high, many trees fall across the channel and remain suspended above it. Trees falling from greater horizontal distances are less likely to be suspended above the channel because their upper trunks shatter or are not supported by the opposite bank. Tree trunks broken into lengths less than or equal to the width of the channel (4 to 5 m) are more likely to enter the effective zone. Trees falling parallel to the channel have a better chance of entering the effective zone because they do not have support points on both banks. Trees succumbing to bank erosion are likely to enter the effective zone. CWD entering the effective zone tends to become pool-associated in these situations: when the piece is massive and is situated such that a pool is formed by scour, if the piece becomes lodged in a debris jam which in turn causes bed-scour and pool formation, or if the piece happens to fall directly into an existing pool. Circumstances in which CWD is associated with pools are limited due to the precise placement required to form pools. The predominance of CWD from A. grandis and P. menziesii reflects successional dynamics of the redwood forest ecosystem. Windthrow, disease, and competition have thinned these species from this 100-year-old secondgrowth forest. The "climax" species, S. sempervirens, is less common as CWD and is particularly uncommon as windthrow. The dominant CWD sources (input mechanisms) are windthrow and bank erosion. Windthrow is the dominant source for the potential zone, much of which originates from adjacent slopes. This is consistent with data for Oregon streams, where 90 percent of CWD originated within 30 m of the stream (McDade 1987). Bank erosion is the most common source of effective zone CWD. Logging debris, mostly in the form of redwood rootwads, is a valuable and stable structural element of the channel, but is a limited resource at Caspar Creek. 169

Discussion And Conclusions

CWD Density The mean density of CWD in Caspar Creek, 339 m3/ha (combined potential and effective zones), is less than that for many streams in old-growth redwood forest. Mean CWD for 12 sample reaches in Redwood National Park were reported by Keller and others (in press) to be 1590 m3/ha. The densities ranged from 240
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Debris jams are major components of the pool-associated class of CWD. Debris jams associated with pools typically are located at channel constrictions, sharp bends, or obstructions (e.g., redwood stumps, boulders). Much CWD in debris jams is small enough to be floated during floods. CWD in the effective zonewhich is likely to contribute to fish habitatoriginates from both far and near the channel and from both the old-growth forest and its successor. Both stable and mobile pieces are associated with pools, which are thought to be the most valuable habitat. Considering the complexity of CWD ecology, one basis for managing CWD is maintaining or increasing the abundance of CWD and the diversity of sources, sizes, and positions.

habitat, indicating that mobile (smaller) CWD is also beneficial.

Acknowledgments
J. A. Baldwin generously provided statistical assistance. L. M. Decker provided fish habitat data for Caspar Creek from an on-going study. Staffs of the Jackson Demonstration State Forest, and California Department of Forestry and Fire Protection provided invaluable assistance and loaned facilities, vehicles, and equipment. R. L. Beschta, M. D. Bryant, and F. J. Swanson provided excellent reviews of the manuscript.

Rate of CWD Accumulation Logging techniques at the turn of the century left little debris in the stream channel at Caspar Creek. Subtracting logging debris from CWD volume, and dividing by 60 years (the first 30 years of regrowth is assumed to have contributed little CWD), yields estimates of annual accumulation rates as follows: Potential & Effective Combined Potential Effective Pool-Associated 5.3 3.5 2.7 1.6 m3/ha/yr m3/ha/yr m3/ha/yr m3/ha/yr

References
Bilby, R.E. 1984. Removal of woody debris may affect stream channel stability. Journal of Forestry 82(10): 609-613. Bisson, P.E.; Bilby, R.E.; Bryant, M.D.; and others. 1987. Large woody debris in forested streams in the Pacific Northwest: past, present, and future. In: Salo, E.O.; Cundy, T.W. eds. Streamside management: forestry and fishery interactions. Contribution No. 57. Seattle: College of Forest Resources, University of Washington: 143-190. Bisson, P.E.; Nielsen, J.L.; Palmason, R.A.; Grove, L.E. 1982. A system of naming habitat types in small streams, with examples of habitat utilization by salmonids during low streamflow. In: Armantrout, N.B., ed. Acquisition and utilization of aquatic habitat inventory information. Portland, OR: Western Division, American Fisheries Society: 62-73. Cochran, W.G. 1977. Sampling techniques. 3d ed. New York: John Wiley & Sons; 413 p. Harmon, M.E.; Franklin, J.F.; Swanson, F.J.; and others. 1986. Ecology of coarse woody debris in temperate ecosystems. Advances in Ecological Research 15: 133-302. Keller, E.A.; MacDonald, A.; Tally, T.; Merritt, N.J. Effects of large organic debris on channel morphology and sediment storage in selected tributaries of Redwood Creek. In: Geomorphic processes and aquatic habitat in the Redwood Creek drainage basin. Prof. Paper. Washington, D.C.: Geological Survey, U.S. Department of the Interior. [In press]. Lienkaemper, G.W.; Swanson F.J. 1987. Dynamics of large woody debris in streams in old-growth Douglas-fir forests. Canadian Journal of Forest Resources 17: 150-156. McDade, M.H. 1987. The source area for coarse woody debris in small streams in western Oregon and Washington. Corvallis: Oregon State Univ. M.S. thesis; 1-69.

For a 100 m reach 5 m wide, the addition each year of one log 4 m long and 0.3 m in diameter would be equivalent to the accumulation rate for the potential and effective zones combined. These estimates could guide efforts to deliberately add CWD to similar streams that are deficient in fish habitat. Professional judgement would be required to determine the appropriate methods and locations needing additional CWD. During logging, unmerchantable logs might be deliberately placed in desired locations. Such additions could be distributed among potential and effective zones to provide an even supply of CWD to the stream over time. Alternatively, if streamside bufferstrips do not preserve an adequate natural source area for CWD, deliberate additions of CWD to the stream might be made in proportion to the expected contribution lost from the source area over the duration of the cutting cycle (e.g., 50 years). Silvicultural methods could also be developed to optimize the supply of natural CWD (Rainville and others 1985). Bilby (1984) and Swanson and others (1984) suggested that massive CWD is more stable and better suited to provide desired fish habitat. We found, however, that CWD in debris jams often contributed to pool

170

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Rainville, R.P.; Rainville, S.C.; Lider, E.L. 1985. Riparian silvicultural strategies for fish habitat emphasis. In: Proceedings of the 1985 Society of American Foresters National Convention; July 28-31; Ft. Collins, CO. Bethesda, MD: Society of American Foresters; 186-196. Swanson, F.J.; Bryant, M.D.; Lienkaemper, G.W.; Sedell, J.R. 1984. Organic debris in small streams, Prince

of Wales Island, southeast Alaska. General Technical Report PNW-166. Portland, OR: Pacific Northwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 1-12. Van Wagner, C.E. 1968. The line intersect method in forest fuel sampling. Forest Science 14: 20-26.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

171

PACIFIC YEW: A FACULTATIVE RIPARIAN CONIFER WITH AN UNCERTAIN FUTURE 1


Stanley Scher and Bert Schwarzschild
2

Abstract: Increasing demands for Pacific yew bark, a source of an anticancer agent, have generated interest in defining the yew resource and in exploring strategies to conserve this species. The distribution, riparian requirements and ecosystem functions of yew populations in coastal and inland forests of northern California are outlined and alternative approaches to conserving this resource are identified. Efforts to obtain additional information on genetic diversity of yew populations and to insure careful management of the species are essential for the protection of this resource.

selective harvesting of the Pacific yew in California, Oregon and Washington. In 1988, 60,000 pounds of bark were harvested from 12,000 trees in western Oregon; another harvest of the same magnitude is scheduled for 1989 (C. Bolsinger, personal communication). This paper briefly outlines the current status of the Pacific yew as a facultative component of riparian forest communities in northern California and other Western states, and addresses questions that bear on possible strategies to conserve the yew resource without intervening with ongoing cancer research.

Yews represent a genetic resource at risk. At least three speciesthe English yew (T. baccata), Florida yew (Taxus floridana), and Pacific yew (T. brevifolia) have evolved an unique assemblage of characteristics that confer upon these trees and shrubs unprecedented vulnerability. The English yew, native to Europe, north Africa and parts of the Middle East, was valued for centuries as a strategic resource. The Florida yew, restricted to bluffs and ravines on the eastern bank of the Apalachicola River in northern Florida (Sargent 1961), is classified as a rare and endangered conifer (Barnes 1983). Now the survival of Pacific yew populations in western states may be threatened by the discovery that yew bark is an important source of an antitumor drug. The yew has the longest tradition as a protected tree. In the 15th century, English yew populations were severely depleted; their fine-grained elastic wood was valued for archery bows. Laws mandating protection of this species date from 1423, (Szafer, W. 1965, cited in Bialobok, 1975). In many parts of central Europe and the Transcaucasus, the English yew remains a rare and vanishing conifer (Karczmarczuk 1986; Safarov 1986). New demands for the yew have recently emerged. Although it has not been listed as an endangered species, environmentalists have expressed concern that populations of Pacific yew in California and other Western states may also suffer depletion (Booth 1987). Yew bark is currently the chief source of taxola diterpenoid antitumor agent (Suffness; Douros 1979; Wani and others 1971). The National Cancer Institute in cooperation with the USDA Forest Service have contracted for the

Pacific Yew as a Riparian Resource


In the Coast Range, on the inland forests of the Klamath and Cascade ranges, and the northern Sierra Nevada of California, the Pacific yew grows as scattered individuals and small groups of trees on or near stream banks and river margins, canyon bottoms and wet shaded ravines, and other moist habitats (Crawford and Johnson 1985; Griffin and Critchfield 1972; Jimerson 1988). Forest inventory data show that this tree is distributed from 100m in the coastal redwood to 1300m in the mixed conifer forests of the central and northern Sierra Nevada. At higher elevations, the Pacific yew may occupy less moist sites on north and east slopes. In the northern part of its range, this facultative riparian species utilizes a variety of moist habitats; however, as it approaches the southern limit of its range, the yew population tends to be more restricted and more obligate in meeting its riparian requirements (Jimerson 1988; F. Johnson, personal communication). Yews are slow-growing, long-lived, and shade tolerant (Minore 1979). In northwestern California, they form an understory layer in association with coast redwood, Douglas-fir, tanoak, bigleaf maple, vine maple, other hardwoods, and conifers (Jimerson 1988). The dense yew subcanopy provides shading; thus, it contributes to maintaining temperature control along tributaries of small streams for juvenile salmon and steelhead (Jimerson 1988). Yews provide cover and serve as an important food source for browsing deer and other ungulates (Pierce 1984; Pierce and Peek 1984). Several

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Adjunct Professor, Department of Biology, School of Environmental Studies, Sonoma State University, Rohnert Park, Calif.; Member, Commission on Education and Training, International Union for the Conservation of Nature and Natural Resources (IUCN). Authors' current mailing address: USDA Forest Service, Berkeley, CA 94701-0245 USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

172

Figure 1 Distribution of Pacific Yew (Taxes brevifolia Nutt.) and location of designated reserve sites containing T. brevolia in northern California.

groups of birdsthrushes, blackbirds, waxwings and nuthatchesare reported to feed on the fleshy aril and help to disseminate seeds of the yew (Bartkowiak 1975; Sudworth 1967).

biodiversity would provide additional data on genetic variation. This information is essential for silvicultural management, for studies of variation in taxol, and for conservation of the species.

The Need for Further Studies


Until recently, the Pacific yew was not considered commercially important; accordingly, virtually no information is available on yew population structure, or how yew trees produce taxol. Many questions need to be addressed, such as: How does the annual growth rate compare with the cutting (harvesting) rate? Studies are needed to provide baseline data on the amount and patterns of genetic variation in the species. Three cultivars of Pacific yew have been reportedErecta, Nana and Nuttallii (Taylor and Taylor 1982). The extensive range of this speciesfrom northern California to southern Alaskasuggests that geographic races have evolved. A range-wide study of isozymes and other measures of
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Conserving the Resource


Gene resources can be preserved in situin reserves, or ex situin seed banks or arboreta (Ledig 1986). Five major natural reserve programs have been established in California. Among the most ambitious are the University of California Natural Reserve System and the USDA Forest Service Research Natural Area program (Ford and Norris 1988). Although Pacific yew is not identified as a target element in established reserves, several contain significant populations of this species (GENREC 1987; Hood 1975-1980) (fig. 1). In outlining the scope of gene conservation for forest trees, Ledig (1988) has underscored the need to establish larger areasgenetic resource management units 173

(GRMUs)to protect coevolved gene complexes. He argues that existing reserves are inadequate; their small size increases the probability of extinction. GRMUs differ from strict reserves; logging can continue, but other forms of management promote genetic conservation. Ledig urges conservationists to evaluate existing reserve natural areas and persuade Federal and State land management agencies to establish GRMUs to fill the gaps. Yew populations need to be preserved in GRMUs now, to assure their future availability. Identifying alternative strategies for producing taxol represents another approach to conserving the Pacific yew. Several methods are being considered: in vitro cell and tissue culture; stimulation of biosynthesis through precursor feeding; and isolation of genes for cloning and expression of taxol. Evidence for cell culture of Pacific yew has only recently been reported from two laboratories (Blume 1989). To date, none of the immediate precursors of taxol have been identified. We know of no studies on the isolation of genes controlling taxol biosynthesis and their expression. Future use of forest genetic resources will be determined by current management strategies and practices. Failure to respond to this challenge could significantly reduce or eliminate the genetic diversity upon which natural selection acts (California Gene Resources Program 1982). One of the victims of that neglect could be the Pacific yew.

Conclusions
The discovery of taxol in yew bark poses a threat to yew populations and a challenge for humankind. We should expect to face such realities in the future as we identify and utilize genetic resources of other plant species. Strategies to both protect and manage the yew resource must be developed quickly if we are to preserve this valuable riparian species.

Acknowledgments
We thank Vincent Dong, Tom Jimerson, and Connie Millar, USDA Forest Service, and Larry Riggs, GENREC for thoughtful reviews and discussion; and Tim Washburn, USDA Forest Service, for generous help in preparing the figure.

References
Barnes, Lee. 1983. Micropropagation of endangered native conifers, Torreya taxifolia and Taxus floridana. Hortscience 18(4): 617. Bartkowiak, Stanislaw. 1975. Seed dispersal by birds. In: The Yew, Taxus baccata L.S. Bialobok, ed. Available from NTIS. Springfield VA.; 139-146 Bialobok, Stefan, ibid.

Summary and Recommendations


The current status of the Pacific yew (Taxus breviifolia Nutt.) as a facultative riparian component of the subcanopy in mixed conifer forests of northern California and other Western States was briefly outlined. Current demands for yew bark as a source of taxola diterpenoid antitumor agentcall attention to the need for additional information on this valuable resource. If clinical research trials with taxol are successful, then pressures to increase the harvesting rate of Pacific yew can be expected. Developing in vitro technologies for producing taxol represents one approach to conserving Pacific yew populations. In addition, range-wide studies of the genetic architecture are essential to provide baseline data for silvicultural management, for variation in taxol, and for conservation of the species. Efforts to identify and preserve a diversity of yew alleles and co-evolved gene complexes must be expanded. Clearly, Pacific yew is in a transitional phase between non-commercial status and commercial exploitation. Careful management of this species will be required to protect this valuable resource.

Blume, E. 1989. Investigators seek to increase taxol supply. Journal of the National Cancer Institute 8(15): 1122-1123. Booth, William. 1987. Combing the Earth for cures to cancer, AIDS. Science 237(4814):969-970. California Gene Resources Program. 1982. Douglas-fir genetic resources. An assessment and plan for California. Berkeley, CA: National Council on Gene Resources; 275p. Crawford, Rex C.; Johnson, Frederic D. 1985. Pacific yew dominance in tall forests, a classification dilemma. Canadian Journal of Botany 63(3): 592-602. Ford, Lawrence D.; Norris, Kenneth S. 1988. The University of California natural reserve system. Bioscience 38(7): 463-470. Genetic Resource Consultants (GENREC). 1987. The California forest genetic sources [FGS] catalog. Version 1.0. User manual. Berkeley, CA: Genetic Resources Consultants; 20 p. Griffin, James R.; Critchfield, William B. 1972. Distribution of forest trees in California. Res. Paper PSW-82. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 114 p.

174

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Hood, Leslie. ed. 1975-1980. Inventory of California natural areas. Sonoma, CA: California Natural Area Coordinating Council; 12 volumes. Jimerson, Thomas M. 1988. Ecological types of the Gasquet Ranger District, Six Rivers National Forest. Forest Service, U.S. Department of Agriculture; 164 p. Unpublished draft supplied by author. Karczmarczuk, R. 1986. English yew Taxus baccata, its history, present status, and its future. Kosmos. Series A. Biology (Warsaw) 35(2): 285-292. Ledig, F. Thomas. 1986. Conservation strategies for forest gene resources. Forest Ecology and Management 14(1): 77-90. Ledig, F. Thomas. 1988. Conservation of diversity in forest trees. BioScience 38(7): 471-479. Minore, Don. 1979. Comparative autecological characteristics of northwestern tree speciesa literature review. Gen. Tech. Rep. PNW-87. Portland, OR: Pacific Northwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 72 p. Pierce, D. John. 1984. Siras moose forage selection in relation to browse availability in north-central Idaho. Canadian Journal of Zoology 62(12): 2404-2409.

Pierce, D. John.; Peek, James D. 1984. Moose habitat use and selection patterns in north-central Idaho. Journal of Wildlife Management 48(4): 1335-1343. Safarov, I.S. 1986. Rare and vanishing species of the eastern transcaucasian dendroflora and their protection. Botanicheskii Zhournal (Leningrad) 71(1): 102-107. Sargent, Charles S. 1961. Manual of trees of North America 1: 94-95. New York: Dover Publ. Co. Sudworth, George B. 1967. Forest trees of the Pacific slope. New York: Dover Publ. Co.; 455 p. Suffness, Matthew; Douros, John. 1979. Drugs of plant origin. Methods of Cancer Research 16: 73-126. Taylor, Roy L.; Taylor, Sylvia. 1981. Taxus brevifolia in British Columbia. Davidsonia 12(4): 89-94. Wani, M.C.; Taylor, L.H.; Wall, M.E. 1971. Plant antitumor agents. VI. The isolation and structure of taxol, a novel antileukemic and antitumor agent from Taxus brevifolia. Journal of the American Chemical Society 93(9): 23252327.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

175

RIPARIAN SYSTEMS AND FOREST MANAGEMENTCHANGES IN HARVESTING TECHNIQUES AND THEIR EFFECTS ON DECOMPOSED GRANITIC SOILS1
John W. Bramhall2 Abstract: In the 1950s, timber on steep granitic terrain in Trinity County, California was harvested by using the logging techniques of the time. After Trinity Dam was built in the 1960s, it became evident these techniques were not suited to quality riparian habitat and healthy anadromous fisheries. Since adoption of the Z'bergNejedly Forest Practice Act in 1973, efforts have been expended to repair damage that has been done to riparian vegetation, and find forest practices compatible with granitic soils. by the many brushfields scattered amongst our existing timberstands.

Effects of Early Logging


Early logging practices were not as regulated as today. Although California's forest practice regulations have always been the most stringent in the nation (Arvola 1976), they have evolved over the years. An example of the evolutionary process is contained in [paragraph 914(c) Logging Practices-Tractor Yarding] in the 1953 edition of the California Forest Practice Rules. Paragraph 914(c), titled "Logging Practices-Tractor Yarding," states in part, "Tractor roads shall be so constructed and left after logging that water flow thereon shall not contribute to undue or excessive erosion of the soils by gullying. Water breaks shall be installed at intervals of not more than three hundred (300) feet or at each natural water course where such courses are less than three hundred (300) feet, on all tractor roads having a gradient greater than ten percent; such water breaks are to be installed immediately following conclusion of use of tractor roads for logging and prior to removal of equipment (State of California 1953)." No mention is made of keeping equipment out of creekbeds. By 1988, the topic of waterbreaks has its own chapter-(Paragraph #934.60), and takes up more than two pages of the California Administrative Code (State of California 1988). Silvicultural practices have evolved in a similar way. Earlier, there was concern about the adequate number of healthy seed trees left to restock the stand (State of California 1953). By 1988 Silvicultural Methods are covered in a separate article (Article #3), spanning ten pages (State of California 1988). No mention is made of clearcutting in the early rules. In 1988, clearcutting is discussed in great detail, including size, shape, width of buffer strips, and successive cutting within the buffers (State of California 1988). Property taxes have also evolved over the years. Earlier, the property tax was levied on the assessed value of the land and the value of the timber. Once a tract of timber had been logged and 70 percent of the volume had been removed, the tract was taxed as

Early logging practices having a major effect on riparian ecosystems included skidding and hauling logs down the creekbeds, harvesting all merchantable trees in the riparian zone, and leaving slash and debris in the creekbeds. There are others, but these are the most identifiable. California passed legislation to regulate forest practices on private land in 1945. A State Board of Forestry was also established at that time. California Forest Practice Rules, created from the legislation, were approved for administration in 1947 (Arvola 1976). Initially there were four foresters assigned to the program as the inspecting cadre. As expected, a few things fell through the cracks. Unfortunately no mention was made of riparian or stream protection zones in these early forest practice rules. By the mid-1960's rules were adopted to protect streams from overzealous construction, substantial diversion, excessive log jams or debris accumulation, and pollution deleterious to fish, plants, or bird life. But pesticides were not named specifically. Contributing to the problem were the remaining attitudes of the earlier timber landowners. One example was the old "cut and get out" philosophy of cutting all of the possible timber in a tract and then moving on to new parcels in other states. Another attitude was exemplified in the belief that a piece of ground would recover soon after it is cut over if it were just left alone. This attitude was left over from operators who were familiar with timber operations in the East. There, where precipitation is plentiful, a piece of land will begin to come back to forest within 5 years if left undisturbed. here, in the arid West, this is not true as witnessed

1Presented 2

at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

Area Forester, Soil Conservation Service, USDA, Red Bluff, California. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

176

"cutover timberland" for the next 40 years (State of California 1964). This method of taxation tended to determine the selected silvicultural system, rather than have that decision based on solid silvicultural principles. In 1977, the State of California adopted a yield tax on forest land (State of California 1987) which taxes the value of the timber yield. Payment is made only at the time of harvest. Yield taxes separate the value of the land from the value of the timber. Land is taxed annually under the property tax while the tax on the timber value is deferred until harvest (Society of American Foresters 1984).

Effects on Granitic Soils


Back in the forties and fifties, logging practices previously described were the same on the granitic soils as those on all other soils. No thought was given to the impact logging operations might have on the granitics. It had not been discovered that granitics react differently to disturbance than do most other soils. The granitic soils are highly erosive. Once they are disturbed, they are extremely difficult to stabilize. Disturbing the stream channel and the adjacent vegetation with equipment for any purpose can cause damage to the riparian zone. This kind of damage in the granitics can be almost impossible to repair. In 1976, the Soil Conservation Service became involved in an emergency erosion and sediment control program in the Grass Valley Creek Watershed (GVC) in Trinity County. GVC is a watershed having about 75 percent granitic soils. Most of the land in the watershed is commercial forest land. The physiography is characterized by steep mountain slopes and narrow valley floors. GVC was managed like other watersheds in the area, but it was different due to the granitic soils. The impact of disturbance started coming to the attention of local officials and residents after the Trinity Dam was constructed in 1963. The dam controlled the peak flows that had ravaged the Trinity River in the past. With the peak flows controlled, the annual spring floods no longer removed the sand deposits in the river. A very large sand bar developed below the mouth of Grass Valley Creek. The Trinity River Task Force, formed in 1973, determined by 1976 that the erosion occurring in GVC was the primary source of the sediment. Part of the erosion control program in GVC has to deal with the riparian zone. The most effective practice appears to be the use of rock rip-rap on the outside bends. This practice is combined with willow plantings to make the project ecologically more sound. Treating streambanks is not the total solution to the problem. Erosion of the roads, landings, and skid trails also contributes to the deterioration of the ecological quality of the stream. Erosion control on these disturbed areas has consisted of the installation of waterbars at a closer than normal spacing; surfacing all permanent roads with rock; down drains and energy dissipaters installed at outlets of drainage ditches and culverts; stream crossings located to minimize disturbance of stream channels and stream flow; and seed, fertilizer, and mulch applied to all fill slopes. Revision of the lake and watercourse classification and protection forest practice rules, adopted in 1983, has made it much easier to prevent additional degradation of the watercourse and lake protection zones (WLPZ).

Recent Forest Practice Legislation

The Z'berg-Nejedly Forest Practice Act of 1973 and its amendments, are the most significant pieces of legislation affecting California's forest resources to date. The riparian resources, considered an integral part of the forest, are also included. In October 1983, new lake and watercourse classification and protection rules took effect. The new rules involved a major change in the procedures for classifying and protecting the waters of the State. Provisions of these rules include clear identification of watercourse and lake protection zones. The width of the protection zones varies with the stream classification and slope class of the adjacent land. There are also rules governing the amount of overstory and understory vegetation that must be left within the protection zone after timber operations (State of California 1983). Although water quality is the primary beneficiary of the watercourse and lake protection rules, the riparian ecosystem is also given major protection (State of California 1988). The rules dealing with erosion control, logging roads and landings, and silvicultural methods, all indirectly enhance the riparian ecosystem. Erosion control serves to keep sediment out of streams and lakes, in addition to keeping soil on the hillside where it belongs. Roads and landings are to be located so that they give adequate protection to water quality and fish and wildlife habitat. Silvicultural methods limit the size of clearcuts and set standards for the use of shelterwood, seed tree, and selection systems; they therefore enhance the riparian ecosystem. These methods help provide or achieve builtin checks for proper forest land management in the form of erosion control, water infiltration enhancement, and maintenance of adjacent non-riparian habitat.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

177

This is true of all soil types, both granitics and others (State of California, 1983). In April 1986, the California Department of Forestry and Fire Protection (CDF) formed a team to develop mitigation measures associated with timber harvesting on decomposed granitic soils in Grass Valley Creek. This team consisted of industrial and agency personnel which met for two days to examine soil erosion problems in decomposed granitic soils associated with timber harvesting. Also, the team was asked to upgrade mitigation measures which go beyond those already contained in the California Forest Practice rules for reducing, to the extent possible, soil erosion associated with timber harvesting on decomposed granitic soils. It drafted nine pages of mitigation measures. Those measures that affect the riparian zone, in this case identical to the WLPZ, are as follows: 1. No new roads will be constructed in the WLPZ unless such locations are explained and justified; and these locations will result in less hazard to the watercourse. 2. Fill slopes that contact stream flow will be armored with competent rock or alternative material that will provide equal or better protection; this shall include the approaches to removed temporary crossings. 3. Permanent culverts will not be used unless the road will be maintained at least annually. Culverts will be sized to a 50-year storm and oversized when streambed conditions indicate. Woody slash and debris must be cleared out for a minimum of 30 feet above the inlet to a culvert. Minimum permanent culvert size is 18 inches in diameter. Use a backhoe for culvert installations and removals except where other methods are justified. Decomposed granitic soils must be compacted in a moist condition when used as fill material around culverts. 4. Place logging slash from road right-of-way along the toe of all fill slopes. 5. Locate landings away from watercourses and outside the WLPZ unless otherwise explained, justified, and approved by the Director of CDF. Flag all watercourse protection zones (WPZ) and equipment exclusion zones (EEZ) prior to harvesting activities. WLPZ width shall be a minimum of 100 feet on all Class I and II watercourses. 6. Increase WPZ and EEZ widths (by at least 50 percent) as necessary to prevent erosion when the following conditions are present: (a) adjacent slopes exceed 50 percent; (b) areas of instability are present; (c) sparse vegetation exists or is projected either prior to or after the operation; (d) excessive natural or human-caused erosion is present; and (e) the water-

course or nearby downstream channels support fisheries. 8. Seed, straw, and fertilize all of the following areas of exposed soil within the WPZ: (a) road (unless rocked), landing, and skid trail running surfaces including sidecast; and (b) temporary crossing approaches. Use brow logs, where needed, to minimize sidecast within the WPZs. The following recommended mitigation measures applying to permanent crossings follow: 1. Minimize excavation by using a backhoe when necessary. 2. Rip-rap inlets and/or install flared inlets. 3. Insofar as is practical, cross watercourses at right angles. 4. Install culverts at the natural channel grade, and control the direction discharge to the grade and center of the stream channel. Temporary crossings had the following recommended mitigation measures: 1. Evaluate the use of metal pipe versus the use of a Humboldt crossing (made of wood) in the Timber Harvest Plan, and explain the reasons for the choice. 2. Cross drainages at right angles, minimize excavations and plan ahead for removal. 3. If a culvert is used, (a) use a backhoe, where practical, for installation and removal; (b) use culverts 18 inches or larger; and (c) use washed rock as fill material if available. 4. If a Humboldt crossing is used, (a) use sound logs just long enough to facilitate skidding or transporting logs, and maximize the number of logs to minimize the amount of fill; (b) when possible, install and remove logs with a front-end loader; (c) preattach chokers prior to installation, or provide for choker attachment upon removal when access restricts using a front-end loader; (d) minimize ground disturbance during installation and removal; (e) keep fill depth to a minimum; and (f) use brow logs on skid crossings where necessary. In conclusion, great strides have been made in protection of riparian vegetation in the eighties. It would be naive to think that by winning a few battles, the war is won. Forty years ago many of the environmental blunders could be charged off to lack of awareness. This is not true any more. Today, all natural resource managers are aware of the impacts management operations have on the resource. The challenge still before us is how to develop mitigation measures that protect the resource, and at the same time, are economically feasible for the operator.

178

SDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

References
Arvola, T. F. 1976. Regulations of Logging in California, 1945- 1975. Sacramento. California, State of California, The Resources Agency, Department of Conservation Division of Forestry, 98 p. Society of American Foresters, 1984. In: Analysis for managerial decision. Karl F. Wenger, editor. Forestry Handbook, Second Edition, Section 17. New York: John Wiley and Sons; 1000-1001. State of California, Division of Forestry. 1953. Forest practice rules for Redwood Forest District. Revised 1953. Sacramento, Calif.; Department of Natural Resources (now, the Department of Conservation); 37 p.

State of California, Division of Forestry. 1964. Forest practice rules for north Sierra Pine Forest District, 1964 Edition, Sacramento, Calif.; Department of Conservation, 51 p. State of California, Department of Forestry. 1983. Guidebook to Board of Forestry watercourse and lake protection rules. Sacramento, Calif.; California Department of Forestry, 90 p. State of California, State Board of Equalization. 1987. California timber yield tax law, 1987 Edition. Sacramento, Calif.: State Board Equalization, Pamphlet No. 43; 80 p. State of California, Department of Forestry. 1988. California Administrative Code, Title 14. Natural Resources, Division 1.5 Department of Forestry. 1988 Revisions. Office of Administrative Hearings, Department of General Services, North Highlands, CA; Articles three, four, and six.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

179

STABILIZATION OF LANDSLIDES FOR THE IMPROVEMENT OF AQUATIC HABITAT 1


Michael J. Furniss2 Abstract: Chronic surface and mass erosion from recent landslides often prevents the recovery of productive stream habitats following initial mass failure events. Low-cost methods that can accelerate recovery and stabilization processes have been employed on numerous failed slopes in the Six Rivers National Forest in the northwest corner of California, with notable success. Two treatment methods, toe erosion control and revegetation have wide applicability in situations where funding is scarce. anything but extremely expensive, large-scale engineering solutions. However, the cost of corrective measures is not always prohibitive, and high value fish and wildlife habitat can often be reclaimed without great expense or effort. Landslide stabilization measures have been applied extensively to road and real estate development, and are well developed. Because of the urgency and high values at risk in these situations, very expensive measures are readily justified and usually include detailed evaluation and engineering designs. Vehicular access is usually feasible. Where wildland water quality or fish habitat are the values at risk, the options for landslide stabilization are usually more limited because of: (1) a lack of perceived urgency and risk - no human life, access or improvements are threatened, (2) small amounts of available funding, often with higher priority projects competing, and (3) lack of vehicular access. Despite these limitations costbeneficial options often do exist and may be an essential step in the reclamation of a stream reach or riparian system.

There is often an important relationship between the condition of the soils in a watershed and the productivity of its riparian areas and streams. The maintenance and improvement of productive riparian systems is largely a consequence of good soil management. In the steep and erodible basins or northern coastal California that means maintaining the stability of slopes and the productivity of their soils. The production of anadromous fish is being emphasized in the coastal streams and rivers on public lands managed by the Forest Service in northern California. In planning the enhancement of fish habitats, it is appropriate to evaluate watershed conditions and identify the factors that limit fish production. On the North Coast of California, landslides are the most common habitat-degrading watershed problem. Large inputs of sediment from landslides often severely degrade or eliminate downstream fish habitat, and create unproductive islands in otherwise productive stream and riparian systems. Large mass failure events are often followed by years of continued surface erosion and secondary mass failures, with each process worsening the other. Chronic sedimentation from subsequent erosion of failed slopes can prevent the natural recovery of downstream channels and riparian systems. It is suggested that such chronic, long-term sediment inputs are much more damaging, volume-for-volume, than catastrophic sediment inputs from mass failure events. The willingness of land managers to attempt any landslide stabilization is often lacking, chiefly because the size of the features creates a perception of futility of
1

Some General Principles


There is a great variety of landslides that can occur in nature resulting from many combinations of causative factors acting on a great variety of landforms. No standard methods of control are universally applicable, but some generalizations about approaches can be made: 1. Corrective measures must be designed to act on the driving forces causing the slide or secondary erosion, or on the resisting forces tending to keep the hillslope stable. The specific factors causing the movement or risk of movement must be identified. 2. Corrective measures will either decrease the destabilizing forces or increase the resistance of the soil or slope to erosion. 3. The consequences of erosional impacts should be evaluated on the basis of the value of the resources threatened. A risk-value versus cost analysis should be done for each possible erosional process and contemplated corrective treatment.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Hydrologist, Six Rivers National Forest, USDA Forest Service, Eureka, California. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

180

The risk is the probability or potential for a particular erosional or hydrologic event to occur. The value is the perceived worth of the resource or beneficial use that would be lost if the event occurred. The combination of risk and value defines the problem (Huffman and Bedrossain, 1979). Corrective treatments will modify the risk and the attendant potential change in value. Comparing riskvalue under various corrective treatments with the cost of corrective treatments facilitates rational decisions as to what to do, and forms a basis for assigning priorities to various processes and treatments. 4 Long-term, self-sustaining treatments that do not require maintenance are preferable to short-term treatments that depend on maintenance for success, particularly in poor access areas. Types of corrective treatments that apply somewhat generally to landslides include: 1. Relocate the affected resource (road, stream). 2. Unload the head of the slope. 3. Load the toe of the slope. 4. Remove water from the slope. 5. Restrain or divert materials. 6. Protect landslide toe from stream erosion. 7. Control gully erosion. 8. Protect erodible surfaces. These same general methods also apply to the prevention of landslides. Detailed discussion of landslide processes and corrective measures can be found in several of the references (Bell 1964; Bradshaw and Chadwick 1980; Transportation Research Board 1978; California Department of Public works 1970; Federal Highway Administration 1978). Site-specific measures are best developed when imagination and judgement are applied and the physical processes involved are understood.

The landslides are known as the Horse Linto, Rib, and Bigtoe Debris Slides. The Rib Debris Slide was initiated, and the Bigtoe and Horse Linto Debris Slides were greatly enlarged by the flood of 1964. The Horse Linto Debris Slide is within the Trinity River watershed and the Rib and Bigtoe Debris Slides are within the Smith River watershed. Each slide had a mass failure volume of several million cubic yards, and each had ongoing severe surface erosion from the barren failure surfaces. The toe of each slide was rapidly being eroded by stream flows. This was a major driving force for ongoing and potential erosion from the slides. Stream gravel sampling above and below each landslide showed substantial increases in both spawning gravel embeddedness and the proportion of fine sediment downstream of each slide, relative to gravels just upstream. Each slide was believed to be adversely affecting high-value habitat for king salmon (Oncorhynchus tshawytscha) and steelhead trout (Salrno gairdneri gairdneri). Each slide was treated in the early 1980's to control toe erosion, and was revegetated to control surface erosion. The objectives were to reduce sediment delivery to affected habitat, hasten the recovery of productive channel conditions, and accelerate the recovery of the soil and ecosystem productivity on the failed slopes. The treatments are being monitored by annual visits for qualitative evaluation, annual photo points, evaluation of the embeddedness of downstream spawning gravels, and survey lines to track slope movement.
Control of Toe Erosion

Three different methods were used to control erosion at the toe of the slides that were being chronically destabilized by erosive river and stream flows: placement of a large prism of rock rip rap; excavation of a new channel with explosives with emplacement of tetrahedron-retard training fences, and; widening of the channel with heavy equipment. On the Horse Linto Debris Slide, an energetic fifthorder stream (Horse Linto Creek) with a narrow channel and steep sideslopes was impinging on the toe of the slide at a sharp meander bend. The stream undercut the slide and promoted its instability, and an active gully system developed on the failure surface. Placement of a prism of rock to protect the toe of the slide was judged to be the best treatment given the circumstances. The channel could not be relocated or modified to reduce the energy of the stream. Access to the toe by heavy equipment was easy to develop, and sufficient funds were available.

Experience With Three Large Debris Slides On The Six Rivers National Forest
The Six Rivers National Forest has treated many landslides that impact anadromous fish habitat. Treatment of three very large slides with common erosional processes and similar resources at risk are described here.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

181

This treatment has been in place for five years and has been completely effective. There has been no toe erosion since the treatment. However, this treatment was quite expensive and this amount of funding is not usually available for such projects. On the Rib Debris Slide, a smaller third-order stream (Idelwild Creek) was eroding the toe of the slide and causing a threat of additional large mass movements as well as ongoing sediment input to the stream from the erosion itself. Vehicular access to the toe of the slide was impractical. A new channel was blasted using serial explosive charges to relocate the stream away from the toe of the slide, and a "retard" was constructed using a line of steel-rail tetrahedrons with cyclone fencing wired to their stream-side faces. This acted as a flow-obstructing energy dissipator, to retard flow velocity in the original channel and cause sediment deposition. This treatment was completely effective. Seven years later, the retard caused extensive sediment deposition and debris accumulation preventing reoccupation of the original channel by the stream. The new channel carries all the flow well away from the toe of the slide. No visible toe erosion has occurred since the treatment. On the Bigtoe Debris Slide, the upper Main Fork of the Smith River (sixth order) was impinging on the toe of the slide, causing ongoing piecemeal mass erosion and promoting surface erosion by removing backed-up sediments. An old rock source development widened the channel somewhat, but riparian vegetation was restricting the effective width of the channel. The River flows are very energetic and funding was quite limited. Placement of rock rip rap would have been appropriate, but was too expensive. Simple widening of the channel and removal of densely growing willows by bulldozer was employed, with emplacement of a small tetrahedron retard along part of the toe. The widened channel is about three times wider than the pre-treatment channel. This spreads high flows, reducing the level of peak flows and effectively reducing their energy. The retards consist of six-foot steel-rail tetrahedrons cabled together. These have accumulated floating woody debris that has increased their effectiveness in dissipating the erosive power of high flows. After two years, it appears that this treatment is effective in substantially reducing but not eliminating toe erosion. The low cost and partial success of this treatment suggests that it is cost-effective where funding is insufficient for rip rap protection and access for heavy equipment exists or is inexpensive to develop.

Revegetation Bare slide surfaces usually are inhospitable to plant growth due to very low nutrient capital and unstable surfaces. They tend to revegetate very slowly. Surface erosion of failure surfaces often removes enormous volumes of sediment before revegetation can stabilize the surfaces. We can accelerate the revegetation and stabilization process through planting and fertilization treatment. Planting and fertilization of slide surfaces, if feasible, is almost always cost-effective. In some situations the newly-established vegetation will simply accompany sediment to the stream or slide toe in mass erosion events. However, the cost of planting and fertilizing is quite low and the long-term benefits are high. The risk of losing the treated areas to additional mass erosion is usually warranted. Many bare and eroding slide surfaces have been planted and otherwise treated for revegetation on the Six Rivers National Forest. The results have been favorable in most places. Vegetative cover, root permeation of unstable soils, and ecosystem productivity have been increased many-fold through simple and inexpensive treatments. While not all slide surfaces are accessible for planting and not all treatments result in successful revegetation, the cost of these treatments and their apparent longterm effectiveness suggests that this can be a highly cost-effective type of treatment. However it is rarely employed, probably because most landslides are believed to be completely untreatable, except with very large amounts of money. We have learned several things from our revegetation efforts on many slide surfaces: 1. Inoculated nitrogen-fixing species such as alder (Alnus sp.), black locust (Robinia pseudoacacia), and sweet clover (Melilotus sp.) should be used because the soils on bare slide faces have not hosted plants before and are therefore very low in nitrogen. Survival rates were not related to nitrogen-fixing, but overall growth and vigor was. Non-nitrogen-fixing species that were planted have not thrived, while those that do fix nitrogen have grown much more quickly and are providing more stabilizing effect. Nitrogen-fixing species will tend to build up nutrient capital in the site, greatly hastening the development of topsoil, plant succession, and ecosystem recovery. 2. Fertilization alone sometimes produces adequate revegetation, with native or local species. In the humid North Coast, seeds of pioneer species are abundant and ubiquitous. If the surface is at least temporarily stable, seedlings will become established each

182

SDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

spring. However, unless there are sufficient nutrients, root growth is too slow to keep up with the summer drought period and seedlings dessicate and die. A single treatment of an N-P-K fertilizer has often proven adequate to achieve revegetation. 3. Cuttings, such as willow (Salix) and coyote brush (Baccaris pilularis) seem inexpensive, but often have a low survival rate. Unless labor is free, cuttings are not inexpensive when the overall success rate of plantings is considered. Our experience is that rooted stock is much more likely to survive and is worth the extra initial cost, if suitable species can be obtained. 4. Vegetation needs no maintenance and gets more effective with time. Deep-rooted woody species require several years to become well-established and effective. Managers must be willing to wait for several years to see distinct beneficial effects.

Landslides that degrade fish habitat and riparian resources can sometimes be stabilized with relatively inexpensive treatments. Land managers should be willing to consider low-cost treatments before concluding that landslides cannot be stabilized.

References
Bell, J. R., 1964. "Mechanics of mass soil movement," In: Proceedings of the 1964 Forest watershed management symposium; Corvallis, Oregon: Oregon State University, page 141-162. Bradshaw, A.D., Chadwick, M.J., 1980 "The restoration of land: The ecology and reclamation of derelict and degraded land," Studies in ecology Volume 6. University of California Press, Berkeley, California. 317 pages. California Department of Public Works, 1970. "Bank and shore protection in California highway practice," Sacramento, California; Division of Highways, State of California, 1970. 423 pages. Huffman, Michael E.; Bedrossian, Trinda L., 1979. The geologist's role in timber harvesting plan review: California Geology, June 1979, California Division of Mines and Geology. 115. Federal Highway Administration, 1978. "Countermeasures for hydraulic problems at bridges - Volume 1 and 2." Report No. FHWA-RD-78-162. 169 pages. Transportation Research Board, 1978. "Landslides - analysis and control," Special Report 176. Robert L. Schuster and Raymond J. Krizek, editors. National Academy of Sciences, Washington, D.C.

Conclusions
In some situations, inexpensive measures for the protection of landslide toes from erosive streamflow can be effective. Channel realignment with flow obstructions to maintain new configurations can be as effective as more costly solutions such as rip rap. Revegetation of bare, eroding slide faces is a costeffective treatment almost anywhere it is feasible, and should be more widely used. Where revegetation is to be employed, nitrogen-fixing species should be included to accelerate soil building. Fertilization can be a useful treatment alone or in combination with introduced vegetation.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

183

SESSION E: COASTAL STREAMS-ECOLOGY AND RECOVERY


European settlement and occupation of California began on the coastal plains, and the earliest impacts of that invasion were concentrated in the riparian zones of the coastal streams. Subsequent urban and residential growth depended on the availability of increasingly far-flung sources of water. Many coastal streams were obliterated, undergrounded, or sucked dry. The balance between the capturing and stabilizing of sediment by vegetation, and the eroding power of freshets has been weighted in favor of erosion. Plant communities formerly well adapted to the hydrologic regime are no longer able to survive, introduced invaders are becoming established, and the animal communities are changing in response. The riparian systems of many coastal streams are unraveling. This session observes the nature of those changes and considers what must be done to mend the balance and heal the streams. Earle W. Cummings State Water Resources Control Board Sacramento, California 95810

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

185

ALLUVIAL SCRUB VEGETATION IN COASTAL SOUTHERN CALIFORNIA 1


Ted L. Hanes, Richard D. Friesen, and Kathy Keane2

Abstract: Certain floodplain systems in southern California sustain a unique scrub vegetation rather than riparian woodlands due to a lack of perennial water. Alluvial scrub occurs on outwash fans and riverine deposits along the coastal side of major mountains of southern California. This vegetation type is adapted to severe floods and erosion, nutrient-poor substrates, and the presence of subsurface moisture. Ten major stands were sampled by line intercepts to determine their species composition, community structure, and successional status. Plant ecology and successional dynamics of these stands are compared. Loss of this unique floodplain vegetation type in the past, and current urban pressures from mining and flood control practices are discussed.

channels exhibit different phases of alluvial scrub vegetation. These phases are related to the amount of time that has elapsed since the most recent flood at each level. Three types of alluvial scrub have been recognized and are related to such factors as the scouring action of flood channels, distance from the flood channel, time since the last catastrophic flood, and substrate features such as texture and moisture content (Smith 1980). The three types can be referred to as: pioneer vegetation is sparse and of low species diversity and stature, and is found within active stream channels or recently scoured streambeds; intermediate - vegetation is rather dense and is composed mainly of subshrubs; and mature - vegetation is composed of fully developed subshrubs and woody shrubs. Mature alluvial scrub is distinguished by its vegetative composition, which contrasts in several respects with that of coastal sage scrub as described by Axelrod (1978), Cooper (1922), Epling and Lewis (1942), Kirkpatrick and Hutchinson (1977), Mooney (1988), Smith (1980), and Westman (1981a,b). Specifically, (1) alluvial scrub has more mesic species than most coastal sage scrub stands; (2) alluvial scrub consists of numerous evergreen shrubs, a diverse assemblage of subshrubs, and a springtime ground cover of annual wildflowers, whereas coastal sage scrub vegetation is composed primarily of drought-deciduous subshrubs with sparse, if any, annual wildflowers; (3) scalebroom (Lepidospartum , quamatum), a shrub with high fidelity to alluvial substrates, is found throughout alluvial scrub communities, but seldom in coastal sage scrub vegetation; (4) species commonly found in chaparral or desert plant assemblages, such as California redberry (Rhamnus crocea), lemonadeberry (Rhus integrifolia), sugarbush (Rhus ovata), mountain mahogany (Cercocarpus betuloides), holly-leaved cherry (Prunus ilicifolia), California juniper (Juniperus californica), and yucca (Yucca whipplei) are also common in the alluvial scrub community, but not in coastal sage scrub vegetation; and (5) small-statured riparian woodland species, such as California sycamore (Platanus racemosa) and mulefat (Baccharis glutinosa) are laced through alluvial scrub stands along major drainages, but are not present in stands of coastal sage scrub.

Scrub vegetation in California is extensive and diverse, occupying coastal and desert dunes, coastal valleys and foothills, interior mountains and desert flats. Holland (1986), in the State's Preliminary Descriptions of the Terrestrial Natural Communities of California, recognizes 73 different scrub types endemic to California. However, one scrub type not included in these descriptions is the unique and threatened vegetation called alluvial scrub. This vegetation is considered a unique habitat with a high priority for preservation by the California Natural Diversity Data Base (1987). Alluvial scrub is an open vegetation adapted to the harsh conditions of the outwash environment. It grows on sandy, rocky alluvia deposited by streams that experience infrequent episodes of severe flooding. This vegetation dominates major outwash fans at the mouths of canyons along the coastal side of the San Gabriel, San Bernardino, and San Jacinto Mountains and lesser floodplain and riverine locations of southern California. Some alluvial scrub species occur also in sandy washes of coastal southern California apart from alluvial fans and large rivers. Alluvial scrub is composed of an assortment of droughtdeciduous subshrubs and large evergreen woody shrubs that are adapted to the porous, low fertility substrate as well as to survival of intense, periodic flooding and erosion. Step-like shrub covered terraces above the wash

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Professor of Botany, California State University, Fullerton, Fullerton, Calif.; Senior Ecologist and Staff Ecologist respectively, Michael Brandman Associates (MBA), Santa Ana, Calif.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

187

Figure 1 Distribution of major stands of alluvial scrub vegetation in southern California, U.S.A.

Methods
Ten stands of alluvial scrub vegetation were studied, comprising the largest, intact stands in southern California. These stands are associated with the Big Tujunga Wash and San Gabriel River (Los Angeles County); Cucamonga, Day, Etiwanda, San Sevaine, Lytle, and Cajon Creeks and the San Antonio and Santa Ana Rivers (San Bernardino County); and the San Jacinto River (Riverside County), (fig. 1). Each stand was sampled by a series of 20 m line intercepts. Several "lines of march" along a compass line perpendicular to the associated stream channel were established within each of the alluvial scrub study sites, beginning at the edge of the stream channel in most cases and progressively moving up onto the higher stream terraces until reaching the site boundary. At 30 m intervals (50 m intervals in degraded alluvial scrub), a 20 m line intercept, perpendicular to the line of march, was established. The intercepts alternated at each interval from right to left of the line of march. Only

subshrubs and evergreen woody shrubs were measured. Plants that were bisected by the intercept line (as projected vertically) were identified by species. Their intercept lengths (to the nearest 5 cm) were recorded, and the number of dead and live individuals of each species was counted along each 20 m intercept. Any intercept length not containing shrub cover was recorded as "bare ground." Intercept data were grouped by development stage of alluvial scrub vegetation (pioneer, intermediate, or mature) for each site. The values derived were used to describe and quantify the structural and successional composition and status of each alluvial scrub site.

Physical Setting
Soils

The floodplain soils upon which alluvial scrub occurs is of two types. Riverwash Association soils are

188

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

found along the main river channels and consist of riverdeposited sands, gravels, cobbles and stones. Inundation occurs each year and is accompanied by scouring, deposition, and removal depending upon the intensity and number of rainstorms each winter. All vegetation is scoured from the main channel during peak flow episodes. Soboba Association soils are located along the terraced banks of the river and are formed of alluvium on the outwash of the rivers. Soboba soils have a cobbly, coarse loamy sand surface underlain by pale brown, single-grain, loosely stratified very gravelly and cobbly sand or loamy sand subsoils. These soils are excessively drained and exhibit very high permeability, very slow runoff, low water-holding capacity and low fertility.

Site Conditions

The sample sites exhibit many similar conditions as well as a few distinctive features. Table 1 summarizes the various site conditions and features.

Results
A total of 237 plant species was encountered on the 10 study sites. To facilitate comparison of species composition between the sites, importance values were calculated for each species. Table 2 lists those species with importance values greater than 5.0 for each site. Eriogonum fasciculatum was an important component of all 10 sites. Lepidospartum squamatum occurred on 9 of the 10 sites. Salvia apiana and Lotus scoparius were next in importance, occurring in 6 of the 10 sites. Artemisia californica occurred on 4 of the 10 sites. Longer-lived, woody species that characterized mature stands (see fig. 2) often had importance values less than 5.0 and are not shown in Table 2. Most (6) of the sites were composed of all three successional stages, but a few sites had two or only one stage. Lytle Creek lacked mature alluvial scrub, whereas San Sevaine lacked intermediate scrub, and Etiwanda lacked both pioneer and mature stages. Presence and absence of developmental stages reflected various factors such as upstream damming or channelization, time since the last major flood, human disturbances, or soil moisture.

Climate

The climate of the study area is Mediterranean, with an annual rainfall of approximately 460 mm, most of which falls during a few heavy winter storms. Once every 10 to 20 years, rainfall far exceeds the norm, resulting in catastrophic floods, such as 1938 and 1969. During such past floods, the rivers have carried debris eroded from upstream slopes and riverbeds, forming broad rocky alluvial outwash fans nearly 600 m thick. The river channels within the fans are subject to frequent scouring and flooding resulting in a barren state. The most significant recent floods in the study area occurred in 1938 and 1969. These and previous major floods produced floodplain terraces bordering the main channels and upon which the three recognizable phases of alluvial scrub vegetation have developed.
1988)

Table 1 - Summary of the Alluvial Scrub Sample Site Conditions and Features (material adapted from Michael Brandman Associates.
Site San Jacinto Mentone Cajon/ Lytle Creeks (Combined) San Sevaine/ Etiwanda/Day Canyons Cucamonga San Antonio San Gabriel Acreage Location 2,677 Upper San Jacinto River 11,440 17,347 12,531 Conditions and Features

3,091

Relatively undisturbed except for small sand and gravel operation, golf course downstream; limited water conservation dikes, diversion channels, and percolation basins. Upper Santa Variously disturbed by rock and sand quarries, aqueducts and railways; industrial, commercial, and residential developments; extensive flood control and water River conservation facilities Cajon and Lytle Crossed by 2 railroad tracks, old Route 66, and (combined) Interstates 15 and Creeks 215. Residential and commercial developments along Route 66 and portions of Cajon and Cable Canyons. Quarrying operations in lower Lytle Creek San Sevaine/ Large lateral flood control dikes, diversion canals, and percolation basin. Major Day Canyon electric transmission corridor and water treatment plant. New extensive residenoutwash tial developments south and west. Cucamonga Flood control structures and percolation basin; small rock quarry. Surrounded by Canyon outwash residential developments.

2,318 San Antonio Streamflow controlled outwash by dam and large lateral dikes. Two highways Canyon outwash cross the site. Quarry operation, small airport and commercial buildings present 300 San Gabriel Streamflow controlled outwash by dams, major lateral dikes, drop structures, and Canyon outwash percolation basins. Crossed by Route 66 and Interstate 210. Paved bikeway on east dike. Major quarrying operations. Dam, recreation area, and county natural area downstream. 1,587 Big Tujunga Natural braided floodplain flanked by hills. Bisected by Interstate 210. Dam, Canyon recreation area, and quarrying operation downstream. outwash

Big Tujunga

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

189

Table 2 - Comparison of Species Composition of Ten Alluvial Scrub Stands Using Importance Values Greater Than 5.0*
Site San Jacinto Dominant Artemisia dracunculus Baccharis glutinosa Croton californica Ericameria palmeri Eriodictyon crassifolium Eriogonum fasciculatum Lepidospartum squamatum Lycium andersonii Opuntia parryi Rhamnus crocea Salvia apiana Adenostoma fasciculatum Baccharis glutinosa Ericameria palmeri Eriodictyon trichocalyx Eriogonum fasciculatum Lepidospartum squamatum Yucca whipplei Eriodictyon trichocalyx Eriogonum fasciculatum Lepidospartum squamatum Lotus scoparius Opuntia littoralis Salvia apiana Toxicodendron divers /ohm Yucca whipplei Artemisia dracunculus Eriodictyon trichocalyx Eriogonum fasciculatum Lepidospartum squamatum Lotus scoparius Opuntia littoralis Opuntia parryi Salvia apiana Adenostoma fasciculatum Artemisia californica Cercocarpus betuloides Eriogonum fasciculatum Lepidospartum squamatum Lotus scoparius Salvia apiana Croton californica Eriogonum fasciculatum Lotus scoparius Salvia apiana Adenostoma fasciculatum Artemisia californica Eriodictyon trichocalyx Eriogonum fasciculatum Lepidospartum squamatum Lotus scoparius Rhamnus crocea Achillea millefolium Artemisia californica Ericameria pinifolia Eriogonum fasciculatum Lepidospartum squamatum Lotus scoparius Malosma (=Rhus) laurina Salvia mellifera Artemisia californica Brickellia californica Eriogonum fasciculatum Lepidospartum squamatum Opuntia littoralis Rhus integrifolia. Baccharis glutinosa Ericameria linearifolia Ericameria pinifolia Eriogonum fasciculatum Lepidospartum squamatum Opuntia parryi Ribes aureum Salix sp. Yucca whipplei Importance Values by Development Stage Pioneer Intermediate Mature 7.74 16.10 7.13 4.96 10.59 35.86 0.94 16.77 0.82 8.02 1.75 31.54 28.84 7.76 0.82 5.38 8.24 84.35 7.40 18.20 51.18 13.10 4.34 4.30 2.25 63.20 30.19 1.10 1.08 1.08 1.08 6.28 15.05 37.35 15.04 10.64 19.40 19.50 17.43 33.93 6.63 18.79 18.87 5.98 12.38 4.69 5.05 3.95 12.46 16.25 8.00 30.98 5.43 5.19 5.48 7.88 6.10 4.13 4.25 11.61 41.17 12.11 7.48 7.38 9.51 28.96 3.91 3.63 16.88 1.23 12.00 8.76 8.88 7.51 7.84 12.18 8.42 16.74

Mentone

Cajon

Lytle Creek

San Sevaine

Etiwanda

Cucamonga

7.54 33.48 9.06 43.70 12.30 30.70 21.09 16.65 7.47 9.50 29.67 10.67 10.09 4.91 6.69 12.99 2.12 22.76 22.53 17.03 8.37 5.08 36.39 15.25 5.31 1.35 10.48

9.12 17.04 15.70 25.96 7.32 1.38 14.08

22.69 37.54 30.46

San Antonio

6.34 23.70 11.91 23.19 11.73 1.88 6.54 8.13 41.25 14.51 15.94 8.13 2.22 11.79 12.19 13.02 22.07 7.76 5.18 33.37 9.72 2.21 5.97 9.85 12.71

San Gabriel

8.10 31.43 29.29 2.41 33.56 13.89 1.36 4.53 3.35

Big Tujunga

*Standard importance values were utilized and divided by 3 to convert them to a base of 100 for convenience.

190

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 2 Profile diagram of alluvial scrub vegetation, showing representation of the three phases of development and their representative species. In order to quantify compositional and structural features of the 3 phases of alluvial scrub at each sample site and to compare the 10 stands, various indices were used: native plant species diversity, percent dominance of the shrub component, and structural diversity of the shrub component (Table 3). These indices were standardized to values between 0.0 and 1.0 against an ideal condition in a manner similar to the Habitat Evaluation Procedure (HEP) developed by the U.S. Fish and Wildlife Service (USFWS 1980). The procedure for establishing and standardizing these indices are currently being documented in a manuscript in preparation by Friesen, Jones, and Keane, and are being referred to as a Habitat Quality Assessment (HQA). San Jacinto, Cajon, and Big Tujunga sites exhibited the greatest species diversity of 8 sites. Etiwanda showed the smallest species diversity, having neither pioneer nor mature stages. Cajon and Cucamonga sites had the highest shrub component of 8 sites with San Jacinto and Big Tujung a sites nearly as high. As with species diversity, the Etiwanda site showed the lowest shrub component. The Cajon and San Gabriel sites exhibited the greatest structural diversity of 8 sites with Etiwanda showing the lowest structural diversity.

Current Status
The distinctive character of alluvial scrub vegetation as it relates to flood-deposited alluvia, makes it one of California's unique riparian systems (fig. 2). The intensity and magnitude of episodic floods creates recognizable vegetation phases that are related to both the age of the stand and to the site conditions of substrate and soil moisture. As a stand develops along a time gradient from pioneer to mature, there is a general trend in species composition replacement from short-lived subshrubs to long-lived woody shrubs. Severe flooding as well as fire and man-caused disturbances can eliminate existing stands of alluvial scrub, and thus initiate new pioneer stands. In contrast, a lack of sufficient soil moisture can prevent an intermediate stage stand from progressing to the mature stage. This condition is best illustrated in the Etiwanda stand. Such intermediate stands may be old in years, but not fully mature in species composition and stature.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

191

Table 3 - Standardized Vegetation Parameters of Ten Alluvial Scrub Sites Based Upon Three Stages of Development Site San Jacinto Mentone Stage of Development Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Pioneer Intermediate Mature Species Diversity1 1.0 0.8 0.9 0.2 1.0 1.0 1.0 0.9 0.9 0.5 Shrub Dominance2 1.0 0.8 0.8 1.0 0.8 0.5 1.0 1.0 1.0 1.0 Structural Diversity3 1.0 0.2 0.4 0.0 0.3 1.0 1.0 0.8 0.4 0.1

Cajon

Lytle Creek

San Sevaine

Etiwanda

0.9 NP

1.0 NP

0.4 NP

Cucamonga San Antonio

San Gabriel

Big Tujunga

0.9 0.7 1.0 NP 0.9 0.8 NP 1.0 1.0 1.0 0.8 1.0

1.0 1.0 1.0 NP 1.0 0.6 NP 0.7 0.8 1.0 1.0 0.7

0.3 0.2 0.3 NP 0.6 0.3 NP 1.0 1.0 0.4 0.3 0.2

1 Species Diversity = D = p i log 2 p 2 where: p i = decimal fraction of total individuals belonging to the ith species

total line intercepts of a shrub species x 100 total length of all intercepts per stage 3 Structural Diversity = ratio of shrubs (greater the 1.0 m tall) to subshrubs (less than 1.0 m tall) * NP = not present
2

Shrub Dominance =

Alluvial scrub vegetation once was more widely distributed along the coastal washes and rivers emanating from the Transverse and Peninsular Ranges of southern California, where coalescing bajadas formed extensive, and in places nearly continuous, skirts along these ranges. Agricultural and urban developments in the past century have resulted in its elimination from most of its former range. As a consequence, alluvial scrub is isolated to stands along unaltered streams and outwashes on major alluvial fans. Industrial and residential developments and flood control projects continue to invade these remaining stands. A number of large projects, including a professional football stadium and an international-class golf course, are proposed for development within some of the premier examples of this vegetation.

Historically, rock and sand mining operations have quarried large pits within the floodplain alluvium upon which alluvial scrub vegetation depends. As these building materials become scarce in the southland, alluvial scrub resources become more threatened. Proposed flood control projects further threaten the integrity and long-term vitality of this unique floodplain scrub type.

Endangered Species
Two alluvial scrub species are listed as endangered by both the U. S. Fish and Wildlife Service and California Department of Fish and Game (State of California 1988): Santa Ana River woolly-star (Eriastrum densi-

192

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

folium ssp. sanctorum) is a much-branched subshrub now restricted to sandy soils on river floodplains or terraced alluvial deposits of the upper Santa Ana River drainage, San Bernardino County. These sites mostly are on privately owned and Bureau of Land Management lands. The Corps of Engineers currently is supporting research on the ecology of the woolly-star. Slender-horned spineflower (Centrostegia leptoceras) is a delicate prostrate annual found on fine-textured, flood deposited river terraces and washes in Los Angeles, Riverside and San Bernardino counties. Extant populations are small and seriously threatened. Most occurrences are on private land and are unprotected.

Future Prospects
At present there is no program of conservation that will ensure the future existence of alluvial scrub vegetation. However, the designation of this vegetation by the California Natural Diversity Data Base (1987) as a unique habitat with a high priority for preservation is a worthy first step. Yet, this designation lacks legal power. The presence of the two rare and endangered plant species in some of the alluvial scrub stands brings both state and federal agencies into action in the protection and management of these alluvial scrub endemics. To a limited extent some alluvial scrub stands are the beneficiary of these agencies' actions. Urban development pressures and flood control structures and practices increasingly will threaten alluvial scrub by direct removal or indirectly by altering the dynamics of its hydrology. Only through the acquisition and management of major stands of alluvial scrub can its future be secured.

References
Axelrod, D. I. 1978. The origin of coastal sage vegetation, Alta and Baja California. American Journal of Botany 65(10):1117-1131. Barbour, M. G. and J. Major, eds. 1988. Terrestrial vegetation of California. 2nd Edition. Sacramento, Calif.: California Native Plant Society, Special Publication No. 9. 1030 p. California Natural Diversity Data Base (CNDDB). 1987. Data base record search for information on threatened, endangered, rare or otherwise sensitive species and communities in the vicinity of the rivers and streams in this study. California Department of Fish and Game, State of California Resources Agency, Sacramento, Calif. Cooper, W. S. 1922. The broad-sclerophyll vegetation of California. An ecological study of chaparral and its related communities. Carnegie Institute of Washington Publication 319. 124 p.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Epling, C. and H. Lewis. 1942. The centers of distribution of the chaparral and coastal sage associations. American Midland Naturalist 27:445-462. Hanes, T. L.. 1976. Vegetation types of the San Gabriel Mountains. In: J. Latting, ed., Plant communities of southern California. Berkeley, Calif.: California Native Plant Society, Special Publication No. 2. Hanes, T. L. 1971. Succession after fire in the chaparral of southern California. Ecological Monographs 41:27-52. Hanes, T. L. 1984. Vegetation of the Santa Ana River and some flood control implications. In: Warner, R. and Hendrix, K., eds. California riparian systemsecology, conservation and productive management. Berkeley, Calif.: University of California Press, 882-888 pp. Hanes, T. L. 1973. The San Gabriel River floodplain, pp 369372. In: Stebbins, G. L. and Taylor, D. W. ,eds., A Survey of the natural history of the south pacific border region, California Biotic themes. Prepared for the National Park Service, U. S. Dept. of Interior, Washington, D.C. Holland, R. F. 1986. Preliminary description of the terrestrial natural communities of California. Sacramento, Calif.: State of California, The Resources Agency, Department of Fish and Game. 156 p., mimeo. Kirkpatrick, J. B. and C. F. Hutchinson. 1977. The community composition of California coastal sage scrub. Vegetation 35:21-33. Michael Brandman Associates, Inc. 1986. Final subsequent environmental impact report: Day Creek land and gravel mining operation and reclamation plan. San Bernardino County, Environmental Public Works Agency. San Bernardino, Calif. Michael Brandman Associates, Inc. 1987. Church Street sand and gravel mining environmental assessment. Prepared for C. L. Pharris Sand and Gravel, Inc., Highland, Calif. Michael Brandman Associates, Inc. 1988. Preliminary draft report, Biological Resources Assessment for Raiders Stadium Project, environmental impact statement/report. Prepared for Lockman and Associates, Monterey Park, Cal. and Department of the Army, Corps of Engineers, Los Angeles District. Los Angeles, Calif. Mooney, H. A. 1988. Southern coastal scrub. In: M. G. Barbour and J. Major, eds., Terrestrial vegetation of California. California Native Plant Society, Special Publication No. 9. Sacramento, Calif. 471-489p. Smith, R. L. 1980. Alluvial scrub vegetation of the San Gabriel River floodplain, California. Madrono 27:126-138. State of California. 1988. 1987 Annual report of the status of California's state listed threatened and endangered plants and animals. The Resources Agency, Dept. of Fish and Game, Sacramento, Calif. 109 pp. mimeo. USFWS. 1980. Ecological services manual - 101 ESM habitat as a basis for environmental assessment. Division of Ecological Services, U.S. Fish and Wildlife Service, Dept. of Interior, Washington, D.C. Westman, W. E. 1981a. Factors influencing the distribution of species of California coastal sage scrub. Ecology 62(2)L:439-455. Westman, W. E. 1981b. Diversity relations and succession in California coastal sage scrub. Ecology 62(1):170-184.

193

RECOVERY OF THE CHAPARRAL RIPARIAN ZONE AFTER WILDFIRE 1


Frank W. Davis, Edward A. Keller, Anuja Parikh, and Joan Florsheim 2 Abstract: After the Wheeler Fire in southern California in July 1985, we monitored sediment deposition and vegetation recovery in a section of the severely burned chaparral riparian zone of the North Fork of Matilija Creek, near Ojai, California. Increased runoff was accompanied by low magnitude debris flows and fluvial transport of gravel, most of which was added to the channel and nearby hillslopes by post-fire dry ravel. The pre-burn riparian forest was dominated by white alder, California sycamore, and coast live oak. Regeneration of these species was entirely by resprouting, due to the absence of local viable seed sources. Recovery of the herb layer was affected strongly by the seeding of Italian ryegrass. Species richness of annuals decreased considerably in the second year, when perennials dominated the riparian zone.

The relatively open understory, mesic soil conditions and the high fuel moisture content of riparian vegetation usually restricts the spread of wildfire through the riparian zone. Flooding is probably a more important recurrent disturbance (Brothers 1985). The combination of intense wildfire and post-fire flooding and sedimentation represents an extreme disturbance of riparian vegetation, and we know of practically no information on the nature of recovery processes after such events. The purpose of this project was to document changes in sediment storage and recovery of riparian vegetation in a chaparral basin after wildfire. A specific objective of the research was to evaluate the effects of wildfire on sediment routing and storage in the chaparral riparian system. Other objectives were to compare the role of seedling recruitment versus sprouting in regeneration of the tree layer, and to evaluate the significance of ryegrass seeding on recovery of the herb layer.

The impact of fire on geomorphic processes is significant in the steep basins of southern California. The dominant sediment transport process on steep chaparral hillslopes is dry ravel or dry sliding of coarse colluvial soil fragments under the force of gravity. Fire increases the rate of dry ravel by removing stabilizing vegetation and litter (Krammes 1965). Rice (1982) reported 1.4 cubic meters per hectare per year of dry ravel in unburned chaparral in southern California, in contrast to 39 cubic meters per hectare for the 3 months following the 1959 Arroyo Seco Fire. Chaparral wildfire can cause high temporary fluvial sediment yields (Scott and Williams 1978). Over 70 percent of the total long term sediment yield occurs in the first year following a fire (Rice 1974). Wells (1987) suggested that debris flows are an important form of sediment transport following fire in small steep watersheds. Keller and others (1988) suggested that large debris flows that affect higher order basins are related to long term instability, with fire and high intensity storms acting as potential trigger mechanisms. Chaparral vegetation reduces flood potential by increasing evapotranspiration, interception, and infiltration (DeBano and others 1979). Vegetation removal by fire reduces evapotranspiration and infiltration. Also, the creation of a water-repellent (hydrophobic) layer in the soil during fire decreases infiltration (DeBano 1974). These post-fire conditions cause an increase in overland flow and streamflow (DeBano and others 1979).

Study Site
We monitored post-fire events in a 270 meter study reach on a tributary to the North Fork of Matilija Creek, a tributary of the Ventura River (fig. 1). The drainage area of the basin is 2.14 square kilometers. Matilija Creek drains the tectonically active San Rafael Mountains, Ventura County, California. The geology of the area is characterized by a sequence of steeply south-dipping Eocene marine sandstones and shales. At least two distinct large debris flow deposits are recognized in the basin. Radiocarbon dating suggests that the oldest deposit (Q2) is 1045 95 BP, and the younger deposit (Q1), sampled in two locations, is approximately 385 85 BP or 295 35 BP (Florsheim 1988). Channel morphology in the study reach is characterized by step-pool sequences, a characteristic bedform in steep channels with coarse, heterogeneous bed material. Detailed data on the pre-burn herbaceous flora in the area were unavailable, but standing trunks of trees that remained after the fire could be identified as a riparian forest of white alder (Alnus rhombifolia Nutt.), California sycamore (Platanus racemosa Nutt.), and coast live oak (Quercus agrifolia Nee. var. agrifolia). Shrubby species included willows (Salix L. spp.), hollyleaf cherry (Prunus ilicifolia (Nutt.) Walp.), laurel sumac (Mal-

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Assistant Professor of Geography, Professor of Geological Sciences, Ph.D. candidate in Geography, Ph.D. in Geological Sciences, University of California, Santa Barbara. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

194

osma laurina Nutt. in T. & G.), poison oak (Toxicodendron diversilobum (T. & G.) Greene.), and Christmas berry (Heteromeles arbutifolia M. Roem.). All species nomenclature in this paper follows Munz (1974), and Abrams and Ferris (1960). A study of spatial distributions and regeneration of riparian tree species in three other canyons (Sisar Canyon, Horn Canyon, and W. Fork Santa Ana Creek) burned in the Wheeler Fire provided supporting data for the research at Matilija Creek (Parikh 1988, fig. 1).

Field Methods
In August 1985, we surveyed a longitudinal profile and 50 cross sections along the study reach (presumably the pre-fire condition), and constructed a contour map of the reach (fig. 2). The longitudinal profile and selected cross sections were resurveyed after the winter storms on January 30-31, 1986, and February 13-15, 1986. Dry ravel deposition was estimated by measuring the width, depth, and length of accumulations and then calculating the volume. Rainfall data for storms that occurred during the winter following the Wheeler Fire were supplied by the Ventura County Department of Public Works. Flow discharge at the study site was estimated using the slopearea method for an ungaged channel (Dalyrymple and Benson 1967).

Figure 2 Contour map of the study reach. The recovery of three riparian tree species and the herb layer in the study reach was documented at monthly intervals from September 1985 to May 1987. Three permanent plots of 250 square meters each were established in the upstream, middle, and downstream parts of the study reach. Sixty-three trees greater than 2.5 centimeters diameter at breast height (dbh) were tagged and mapped, including 23 alders, 19 sycamores, and 21 oaks. These trees were monitored to determine rates of resprouting, as well as the fate of dead standing timber in the riparian zone. Ten seed traps were located in each permanent plot to estimate monthly seedfall from alders and sycamores. Greenhouse germination tests for seed viability were also conducted on seeds collected at the site. Thirty plots of 1 square meter each were set up on different geomorphic locations to monitor the recovery of herbaceous species. The latter included Italian ryegrass (Lolium perenne L. ssp. multiflorum (Lam.) Husnot.), which was seeded in the area by the Forest Service, U.S. Department of Agriculture, for erosion control at an average density of 230 seeds per square meter.

Figure 1 Locations of study sites.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

195

Results
Effect of the Wheeler Fire on Hillslope Erosion

The Wheeler Fire burned 480 square kilometers in Los Padres National Forest, California in July 1985. Dry ravel was the dominant erosion process that occurred on burned slopes in the drainage basin following the fire. Slopes throughout the basin and in all lithologic units contributed dry ravel to the channel. Clasts averaged 4 millimeters in diameter. The volume of gravel which entered the channel in the study reach by the process of dry ravel was measured in August 1985 as 0.20 cubic meters per meter length of channel (it is assumed that this deposition post-dates the fire). The volume of dry ravel accumulation measured in a 176 meter reach in the burned headwaters of the North Fork of Matilija Creek near the study basin was 200 cubic meters per hectare (0.30 cubic meters per meter length of channel) for a 7 month period following the fire. Thus, we estimate the rate of dry ravel following the fire in the basin to be 29 cubic meters per hectare per month, much higher than the background rate suggested by Rice (1982).
Effect of the January 30-31 Storm

curred in the constrictions. The distribution of sediment deposits resulting from this low magnitude event is similar to that predicted using the HEC-2 step-backwater computation to model the distribution of unit stream power (Florsheim and Keller 1987). The model predicted that sediment should be deposited upstream of constrictions where backwater effects were present, and downstream of constrictions in channel expansions, leaving constrictions free of sediment.
Effect of the February 13-15 Storm

The second storm, resulting from 244 millimeters of precipitation (return period 2-4 years), occurred on February 13-15 and produced a flow discharge of 2.5 cubic meters per second, only slightly higher than the January 30-31 flow. This event scoured nearly all the gravel deposited by the late January event, suggesting that this flow was supply-limited. The February flow was capable of transporting the small gravel out of the system, and because slopes were depleted and little new sediment was added to the channel in the 2 weeks between the January and February storms, the later flow nearly eroded the channel to the level of its pre-January thalweg.

Two storms caused significant channel change the winter following the Wheeler Fire. The first storm, on January 30-31, 1986, resulted from 122 millimeters of precipitation (return period less than 2 years) and produced a flow of 2.1 cubic meters per second in the study reach. This was the first substantial flow following the Wheeler Fire, and the basin responded to this moderate sized event by contributing material from hillslopes and mobilizing sediment existing in the channel. The streamflow was clearly transport-limited and deposited fine-grained gravel-sized sediment, which filled pools and buried bedforms. Approximately 550 cubic meters of fine-grained gravel was deposited in the 270-meter study reach. Based on a comparison of the volume of gravel-sized material deposited during the January storm (approximately 2 cubic meters per meter length of channel) to the volume of dry ravel existing in the channel following the fire but before the storm, only 10 percent of the gravel-sized material in the January fill was derived from gravel existing in the channel. The remainder was derived from fine material stored on hillslopes near the channel and contributed to the channel by rills and small surficial slope failures during the storm. The pattern of sediment deposition during the January event was dependent on channel geometry. Sediment filled the channel both upstream and downstream of bedrock constrictions, completely burying the prestorm bed morphology (fig. 3). Little aggradation oc-

Figure 3 Longitudinal profile of lower portion of the study reach. Bed profile surveyed in August 1985; following the January 30-31, 1986 event; and following the February 13-15, 1986 event. Constriction extends from approximately 170-200 meters.

196

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

per 1000 years. This calculation suggests that fire may cause 30 to 50 percent of the total debris deposition over a 1000-year period. The remaining deposition is due to high magnitude debris flows related to factors such as basin stability and intensity of precipitation, with fire and earthquakes acting as potential trigger mechanisms. Small gravel deposited in the channel is stored for short periods of time, on the order of months or several years. However, large debris flows remain in the channel as terraces and lag deposits for hundreds to thousands of years.
Vegetation Recovery

Figure 4 Cross section 6 surveyed in August 1985 and in June 1986 shows January fill in the left channel and scour resulting from the February storm in the right channel. Pronounced changes in channel morphology occurred at the downstream end of the study reach during the February event. Sediment stored in the channel was eroded during the February flow and bank erosion occurred locally along vertical edges of the debris flow terrace. The channel widened 0.5 meters and deepened 0.7 meters at cross section 6 as a result of erosion during the February flow (fig. 4).

Tree Overstory Layer

Importance of Fire on Long Term Sediment Storage and Yield

The impact of fire on long term sediment storage and yield can be estimated by comparing the volume of sediment deposited during sediment transport events such as the January 30-31 storm to deposits of large debris flows preserved in the channel. The volume of sediment deposited in debris flows Q1 and Q2 before subsequent incision is estimated as 32,800 cubic meters and 65,600 cubic meters, respectively. Thus, the total volume of sediment deposited in debris flows in the study basin in the past 1000 years is approximately 94,000 cubic meters. The average volume of sediment deposited during the January 30-31 storm was approximately 2 cubic meters per meter length of channel. Assuming that the upper 850 meters of the basin was too steep to allow for sediment deposition, and that sediment was transported to the lower part of the basin (1640 meters upstream of the mouth), the volume of sediment deposited in the first substantial post-fire storm was 3,280 cubic meters. Assuming that the recurrence interval for fire in the Santa Barbara area is 30-65 years (Byrne 1979), and that a sediment transport event such as that on January 3031 occurs following every fire, then 15-33 such events occur in 1000 years. Therefore, the volume of sediment contributed to the basin due to post-fire fluvial sediment transport is estimated as 49,200 to 108,200 cubic meters

The riparian zone of North Fork Matilija Creek was dominated by a central corridor of white alders, with California sycamores and coast live oaks situated at higher elevations above the stream on upper terraces and hillslopes. These relationships were typical of the watersheds in the region (Parikh 1988). There was no apparent relationship between the spatial and sizeclass distributions of the three species. The range of occurrence of the three species above the channel is given in table 1. Most tree trunks along the channel reach remained standing until the storm events in winter 1986, when large flows caused the erosion of mid-channel bars, and subsequent uprooting of many alders growing near the stream. In the first year after the fire, 8 of 23 alders fell, and 2 more were uprooted in the next 2 years (i.e., 44 percent of alders in the study reach, and 3.84 percent of the total basal area of alders on that plot). No sycamores fell during the first 2 years after the fire. During the second year, 6 of 19 oak trees fell following severe windstorms (i.e., 29 percent of oaks in the study reach, 11.04 percent of the total basal area of oaks on that plot). The combined effect of wind and rain during winter storms in the second year caused much breaking of dead tree trunks and branches, and the accumulation of debris in the wider parts of the channel. Table 1 Elevations of species in meters above the channel, North Fork Matilija Creek Range of Occurrence 0.05 to 1.32 0.41 to 4.76 1.29 to 7.80 Standard Mean elevation deviation 0.51 2.84 4.15 0.33 1.11 2.63

Species White alder California sycamore Coast live oak

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

197

The regeneration of alders, sycamores, and oaks occurred exclusively by sprouting. Seedlings were not observed in the permanent plots, although they occurred in unburned areas immediately downstream of the study reach. Less than 1 percent of alder seeds collected from the burned areas were viable. Alder seedfall was consistently low (< 7 per square meter) during the study period. Burned sycamores released large numbers of seeds (up to 170 per square meter) immediately after the fire, but viability was low (< 5 percent). The maximum number of seeds fell about 6 months after the fire (late fall and early winter are the seasons of seed dissemination of both species). Among the trees surveyed in May 1987, 7 percent of alders had sprouted, 83 percent of the sycamores, and 70 percent of the oaks. Surviving alders sprouted several months after the other species. At the study site and in sample transects located in three other burned canyons in the Ojai area (Parikh 1988, fig. 1), alders sprouted the least, and sycamores the most. Alder seedlings were more numerous than sycamore seedlings along the unburned channel in the second year. In the third year after the fire, some of these alder seedlings grew rapidly during the spring season, reaching heights of almost 2 meters; others were flooded out during preceding winter storms. Stepwise logistic regression analysis (Ezcurra and Montana 1984) was used to test the rate of re-sprouting by each species as a function of canyon (N. Fork Matilija Creek, W. Fork Santa Ana Creek, Sisar and Horn Canyons), size-class (dbh), and height above the stream (Parikh 1988). The sprouting of alders was affected most by canyon location (p < 0.0001). The sprouting of sycamores was negatively and significantly related to height above the stream (p < 0.005) and secondarily to regional location (p < 0.1). The sprouting of oaks was positively and significantly related to size class (p < 0.0001). The differences between canyons in rate of resprouting by alders and sycamores is probably due to regional variation in fire intensity. Height above the stream is a factor operating at a smaller scale. More sycamores sprouted closer to the stream, probably due to fire being less intense in the mesic areas of the canyon, and more water being available for sprouting. The thick bark of larger oaks probably accounts for their higher survival rate (Plumb 1980).
Herb Layer

We began monitoring the herb layer in September 1985. Fifteen annuals occurred in the plots during the first year after the fire, the most frequent of which are listed in table 2. Some of these were fire-followers (e.g., Eucrypta chrysanthemifolia, see Keeley and others 1985). The second-year annual flora decreased in species richness, and annual species present in both years generally showed considerably reduced frequencies in the

second year. Most annuals occurred patchily on higher geomorphic surfaces such as hillslopes and upper terraces (fig. 5). Sanicula crassicaulis and Dichelostemma pulchella also occurred on higher geomorphic areas in the first year after the fire. Perennials such as Toxicodendron diversilobum, Rubus ursinus, and Calystegia macrostegia were common in the entire riparian zone, while others such as Solanum douglasii, Artemisia douglasiana, Urtica holosericea and Mimulus cardinalis were more frequent on lower and middle terraces (fig. 5). Several shrub species developed and persisted after the burn on various geomorphic surfaces, e.g., Malosma laurina, Eriophyllum confertiflorum, and Prunus ilicifolia. A number of species were flooded out in winter storms during the first year. Some species, viz Artemisia douglasiana and Solanum douglasii, did not recover fully from these storms until the second year. The second-year perennial flora was somewhat different in species composition. Although Toxicodendron diversilobum, Rubus ursinus, Urtica holosericea, and Solanum douglasii persisted, species such as Baccharis glutinosa, Heteromeles arbutifolia, and Eriogonum fasciculatum ssp. foliolosum normally components of undisturbed riparian, chaparral, and coastal sage scrub communities also appeared. Woody perennials dominated the lower riparian zone in the second year, notably willows (Salix L. spp.). Ryegrass that was seeded in the area in October 1985 germinated in December. Sedimentation events that followed the fire affected the spatial distribution of the species considerably it grew most densely on lower hillslopes and on terraces where sediment accumulated due to dry ravel processes (fig. 6). Maximum density occurred soon after germination (January 1986) and declined thereafter due to thinning and, on lower terraces, due to mortality during flooding (fig. 6). The ryegrass attained an average frequency of 35 percent and average density of 91 plants per square meter (i.e., 39 percent recruitment from broadcasted seed) in the first year. Second year growth started in November 1986, and the species occurred at 50-100 percent frequency on all geomorphic surfaces. At the end of the first year growth in spring 1986, the subplots were harvested, and dry biomass was measured for all herbaceous species on each subplot (fig. 7). Ryegrass dominated the plant communities at all locations, suggesting that native plants were reduced considerably by the growth of this species (Nadkarni and Odion 1986).

198

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 2 Frequent post-fire species observed (*) at North Fork Matilija Creek
Scientific Name Common Name Abbre- Year 1 Year 2 viation

Annuals and Biennials Allophyllum divaricatum (Nutt.) A. & V. Grant Claytonia perfoliata Donn. var perfoliata Collinsia concolor Greene. Dichelostemma pulchella (Salisb.) Heller. Eucrypta chrysanthemifolia (Benth.) Greene var. chrysanthemifolia Callum aparine L. Gnaphalium californicum DC. Lolium perenne L. ssp. multiflorum (Lam.) Husnot Phacelia cicutaria Greene var. hispida (Gray) J.T. Howell Phacelia viscida (Benth.) Ton. Pholistoma auritum (Lindl.) Lilja. Rafinesquia californica Nutt. Sanicula crassicaulis Poepp. ex DC. Sticky phacelia Common fiesta flower California chicory Pacific sanicle Pv Pa Rc Sc * * * * * * * Caterpillar phacelia Pc * * Goose grass Green everlasting Italian ryegrass Ga Gc Lp * * * * * Miner's lettuce Southern chinese houses Wild hyacinth Common Eucrypta Cp Cc Dp Ec * * * * * * * * Straggling gilia Ad * *

Perennials Artemisia douglasiana Bess in Hook. Baccharis glutinosa Pers. Calystegia macrostegia (Greene) Brummitt ssp. cyclostegia (House) Brummitt Eriodictyon crassifolium Benth. var denudatum Abrams. Eriogonum fasciculatum Benth. ssp. foliolosum (Nutt.) Stokes Eriophyllum confertiflorum DC. Gray var. confertiflorum Heteromeles arbutifolia M. Roem. Lotus scoparius (Nutt. in T. & G.) Ottley ssp. scoparius Malosma laurina Nutt. in T. & G. Mimulus cardinalis Dougl. ex Benth. Mimulus longiflorus (Nutt.) Grant ssp. longiflorus Phacelia ramosissima Dougl. ex Lehm. var. suffrutescens Parry. Prunus ilicifolia (Nutt.) Walp. Rubus ursinus C. & S. Salvia mellifera Greene. Solanum douglasii Dunal in DC. Urtica holosericea Nutt. Holly-leaved cherry California blackberry Black sage Douglas' nightshade Hoary nettle Pi Ru Sm Sd Td Uh * * * * * * * * * * * * Branching phacelia Pr * Laurel sumac Scarlet monkey-flower Salmon bush monkey-flower MI Mc Mc * * * * * Christmas berry California broom Ha Ls * * Golden yarrow Ecc * * Thick leaved yerba santa California buckwheat Ef Ecd * Douglas' mugwort Mule fat Coast morning-glory Ad Bg Cm * * * * *

Toxicodendron diversilobum (T. & G.) Greene Poison Oak

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

199

Figure 5 N. Fork Matilija Creek: post-fire development of species. Largest square represents maximum species frequency of 100 percent. 1,2,3 = lower, middle, upper terrace; 4,5,6 = lower, middle, upper hillslope. For abbreviations of species names, see table 2.

The study reach was monitored at quarterly intervals during the third year after the fire. A final survey of the site in spring 1988 (about 3 years after the burn) showed some changes in the flora. Most subplots were dominated by perennials such as Rubus ursinus, Urtica holosericea, Baccharis glutinosa, Calystegia macrostegia, Mimulus cardinalis, Toxicodendron diversilobum, and Salix spp. Shrubby chaparral species that persisted from the first 2 years included Malosma laurina, Heteromeles arbutifolia, Prunus ilicifolia, Eriogonum fasciculatum, Eriophyllum confertiflorum, and Salvia mellifera. Some new species were seen, for example, California lilac (Ceanothus L. sp.) The most interesting feature of the third-year postfire flora was a return to the species composition of the first year on some subplots. Species included Phacelia cicutaria, Eucrypta chrysanthemifolia, Pholistoma auritum, Allophyllum divaricatum, and Galium aparine. Ryegrass was frequent on most subplots. New species observed in June 1988 downstream of the study site in an area burned in August 1987 included short-lobed phacelia (Phacelia brachyloba) (Benth.) Gray), west-

ern morning-glory (Calystegia purpurata (Greene) Brumitt.,) and iris-leaved rush (Juncus xiphioides) E. Mey.)

Discussion and Conclusions


Hydrologic processes in drainage basins in southern California are controlled by infrequent storms and episodic streamflow, and erosional processes are dominated by frequent dry season gravitational sliding of small gravel sized clasts (dry ravel) and infrequent large magnitude debris flows. Hillslope stability and fluvial processes in chaparral basins are influenced by periodic fire.

200

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 6 Matilija Creek: post-fire density of seeded ryegrass. Largest square scaled to represent maximum ryegrass density of 310 plants per square meter. 1,2,3 = lower, middle, upper terrace; 4,5,6 = lower, middle, upper hillslope.

Fire decreases basin stability; sediment deposition during the January 30-31, 1986, storm may be typical of the chaparral ecosystem following wildfire. Evaluation of debris flow deposits suggests that the recurrence interval of large debris flows in a particular basin is at least hundreds and perhaps thousands of years. By comparison, the recurrence interval of wildfire in the Santa Barbara area is 30-65 years (Bryne 1979). Thus fluvial transport of sediment derived from dry ravel and small landslides off hillslopes is much more common following a fire than is a large debris flow. The recovery of vegetation following chaparral wildfire shows the combined disturbance effects of fire and flooding. Recovery processes in the herb layer are closely linked to geomorphic location in the riparian zone, and to the density of seeded ryegrass. Annuals become wellestablished on higher geomorphic locations less prone to flooding, but often in loose soil subject to dry ravel. Perennials, on the other hand, grow better on lower, more disturbed geomorphic locations near the stream. The overall species richness of annals decreased in the second year after the fire due to the predominance of ryegrass, although perennials took over the riparian zone to a large extent.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Sprouting is the dominant means of recovery of the tree species due to the lack of viable seeds following the fire. Certain species such as sycamore show rapid recovery, while others such as alder may be very slow to reestablish in the absence of a viable seed source. Full recovery of the alder canopy after unusually hot fires such as the Wheeler Fire may take many years or decades.

Acknowledgements
We thank personnel of the Forest Service, U. S. Department of Agriculture (Ojai Ranger District Office and the Supervisor's Office of the Los Padres National Forest in Goleta), especially Jim O'Hare, for logistical support during the study. We also thank Elva Rogers, Julia Boles, Ed Harvey, Kathryn Thomas, Nathan Gale, (and other students in the Geography and Biological Sciences departments at University of California, Santa Barbara, for field assistance; Susan Webb and Shirley Brous, Department of Geography, for typing the manuscript; and Janice LaTuchie, for cartographic assistance. This study was supported by Agency Award number W-687, Water Resources Center. University of California. Davis. 201

Figure 7 Matilija Creek: dry biomass of herbaceous species harvested June 2, 1986. 1, 2, 3 = lower, middle, upper terrace; 4, 5, 6 = lower, middle, upper hillslope.

References
Abrams, L.; Ferris, R. 1960. Illustrated flora of the Pacific states. Vol. I-IV. Stanford, CA.: Stanford University Press. Brothers, T. S. 1985. Riparian species distributions in relation to stream dynamics, San Gabriel River, CA. University of California; Los Angeles: Ph.D. dissertation. 120 p. Byrne, R. 1979. Fossil charcoal from varved sediments in the Santa Barbara Channel: an index of wildfire frequencies in the Los Padres National Forest (735 A.D. to 1520 A.D.). A report submitted to the USDA Forest ServiceUniversity of California, Berkeley, 69 p. Dalyrymple, T.; Benson, M. 1967. Measurement of peak discharge by the slope-area method. Techniques of water resources investigations of the U.S. Geological Survey, A2, 12 p. DeBano, L.F. 1974. Chaparral soils. In: Symposium on living with the chaparral, Proceedings; March 30-31, 1973; Riverside, CA. Sierra Club: 19-26. DeBano, L.F.; Rice, R.M.; Conrad, C.E. 1979. Soil heating in chaparral fires; effects on soil properties, plant nutrients, erosion and runoff. Res. Paper PSW-145. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station,

Forest Service, U.S. Department of Agriculture; 21 p. Ezcurra, E.; Montana, C. 1984. On the measurement of association between plant species and environmental variables. Acta Oecologica/Oecologica Generalis 5: 2133. Florsheim, J.F., 1988. Channel form and process: A modeling approach. University of California; Santa Barbara: Ph.D. dissertation, August 1988, 170 p. Florsheim, J.L.; Keller, E.A. 1987. Effect of fire and channel morphology on fluvial sedimentation. In: Erosion and sedimentation in the Pacific Rim. IAHS Publ. no. 165; 279-280. Keeley, J.; Morton, B. A.; Pedrosa, A.; Trotter, P. 1985. Role of allelopathy, heat and charred wood in the germination of chaparral herbs and suffrutescents. Journal of Ecology 73: 445-458. Keller, E.A.; Florsheim, J.L.; Best, D.W. 1988. Debris flows in the chaparral: myth and reality. GSA Abstracts with Programs 20(3): 172. Krammes, J.S. 1965. Seasonal debris movement from steep mountainside slopes in southern California. In: Proceedings of the Federal Inter-agency Sedimentation Conference, Jackson, MS. U.S. Department of Agriculture, Misc. Pub. 970; 85-88. Munz, P. 1974. A flora of southern California. Berkeley: University of California Press; 1086 p.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

202

Nadkarni, N.; Odion, D. 1986. Effects of seeding an exotic grass Lolium multiflorum on native seedling regeneration following fire in a chaparral community. In: Proceedings of the Chaparral Ecosystems Research Conference;1985, May 16-17; Santa Barbara, CA. Report no. 62. Davis: California Water Resources Center, University of California, Davis; 115-121. Parikh, A. 1988. Factors affecting the distribution of riparian tree species in southern California chaparral watersheds. Department of Geography, University of California, Santa Barbara: Ph.D. dissertation draft. Plumb, T. R. 1980. Response of oaks to fire. In: Proceedings of the Symposium on the ecology, management and utilization of California oaks. General Technical Report PSW-44. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U. S. Department of Agriculture; 216-219 p.

Rice, R.M., 1974. The hydrology of chaparral soils. In: Symposium on living with the chaparral, Proceedings: March 30-31, 1973; Riverside, CA. Sierra Club:27-34. Rice, R.M. 1982. Sedimentation in the chaparral: How do you handle unusual events? In: sediment budgets and routing in forested drainage basins. Swanson, Frederick J; Janda, Richard J.; Dunne, Thomas; Swanston, Douglas N. eds. 39-49 p. Scott, K.M.; Williams, R.P. 1978. Erosion and sediment yields in the Transverse Ranges, southern California. U.S. Geological Service Professional Paper 1030, 37 p. Wells, W.G. 1987. The effects of fire on the generation of Debris flows in southern California. In: Debris flows/ avalanches: processes, recognition, and mitigation. Costa, J.E.; Wieczorek, G.F. eds. Reviews in Engineering Geology 7:105-114.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

203

RIPARIAN RESTORATION AND WATERSHED MANAGEMENT: SOME EXAMPLES FROM THE CALIFORNIA COAST 1
Laurel Marcus2 Abstract: Managing and restoring watersheds often involves re-creation of riparian habitats. The natural functions of riparian forest natural to slow flood water, stabilize stream banks and trap sediments can be used in restoring disturbed creek systems. The State Coastal Conservancy's wetland enhancement program is preserving wetlands on the California coast through repair and management of watersheds. Examples from rural and urban areas illustrate how riparian restoration can be integrated into watershed projects. The lagoon originally was a tidal system which probably closed off from the tides during the summer months. The beach berm at the lagoon mouth, remained in its natural state allowing the lagoon to open and flush out sediment during winter floods. During the 1940's the basic hydrology of the lagoon changed housing was developed on the beach berm, the lagoon mouth location was fixed and a weir was placed at the mouth. The weir created a freshwater lagoon and maintained a consistent water level in the lagoon for aesthetic reasons. Unfortunately these changes, combined with the construction of three road crossings in the lagoon, created a perfect settling basin for sediment. Concurrent with these changes in the lagoon, the watershed was dramatically altered. The mainstem of Buena Vista Creek has perennial waterflow making an environment suited for dense riparian growth. The natural floodplain served to slow floodwaters and sieve out heavy sediments. The floodplain gradually builds up sediment protecting the lagoon from rapid filling. During the 1950's 70's much of the Buena Vista Creek floodplain was developed with a primary focus on creek "improvements" concrete channels to drain water as quickly as possible off the development site. These channels not only have no sediment retention qualities, they serve to greatly accelerate stormwater flow rates. In addition to altering the creek, development resulted in the paving of many square miles of the lagoon watershed. The overall effect of both changes was an enormous increase in the peak volume of stormwater and a loss of natural floodplain. This dramatic transformation altered the hydrologic balance of the watershed. The increased volume of stormwater eroded a 20 foot gully in the channel toppling riparian trees and moving tons of sediment downstream. As the stream system attempts to regain a balance with this increased water volume, this gully will widen and deepen causing trees to be undercut and fall over. An arroyo will eventually form particularly if development progresses (Figure 1).

The coastal wetlands of California have dwindled in acreage to one-fourth of their former extent. While much of this loss has come from diking, filling, and dredging development, indirect filling from sedimentation is a major problem. With the passage of the California Coastal Act in 1976, the direct filling of wetlands became severely restricted. However, indirect filling is a largely unregulated and ubiquitous problem involving large land areas, many government jurisdictions and all sorts of land uses. What does the sedimentation of coastal marshes have to do with riparian habitats? There is a very direct relationship. The California State Coastal Conservancy through its efforts to restore and preserve coastal wetlands has had to look upstream into watersheds and stream systems to find the sources of sediment and reduce erosion problems. The following two examples demonstrate how the preservation and restoration of riparian systems helps protect coastal marshes and estuaries.

Buena Vista Lagoon


The Problem

Buena Vista Lagoon lies in northern San Diego County. Its small 20 square mile watershed is linear in shape; the upper half lies in the city of Vista and the lower half, as well as the lagoon, lie in the cities of Oceanside and Carlsbad.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Project Analyst, California State Coastal Conservancy, Oakland, California. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

204

Figure 1- Arroyo Formation.


USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

205

In 1978-79, a series of large storms filled the eastern basin of the lagoon with sediment and brought an outcry from local citizens. The California Department of Fish and Game which owns the lagoon spent over one million dollars to dredge out the muck; however, the lagoon still retains much of the sediment it received.
Sediment Control Plan

Sedimentation-periodically dredge it from the lagoon, or find and correct the source of the erosion. The Conservancy chose the second option and has thus become involved in riparian habitat restoration. The primary reason for this choice lies in a rule of sediment transport/the quantity of sediment transported in a creek is an exponential function of the quantity of stormwater flow. Consequently, the greatest control of sediment into Buena Vista Lagoon can be achieved by reducing peak flows of stormwater. This concept was the product of a comprehensive watershed plan prepared by Applegate and Associates that involved extensive computer modeling and evaluation of cost-effective control measures. The Conservancy plan recommends two primary measures to reduce the cause of sediment movement-urban stormwater flows: detention of stormwater, and creekrestoration.
Stormwater Detention

areas, the channel dimensions must be reduced and the side slopes made relatively steep (2:1 and 3:1). The installation of the detention basins is an integral part of the creek restoration. The flood flows would be lowered enough by the basins to allow for a fully vegetated channel in a confined area (Figure 2). The width restriction requires use of dense plantings of tree species over the entire slope using an irrigation system. The engineering design also incorporates a high channel roughness coefficient so that the channel bottom may be planted and maintained with tree species. Channel maintenance will involve hand cutting of trees over a certain diameter; no dragline work in the channel is envisioned. While this maintenance scheme may seem excessive, it is the only way to have the channel continuously slow stormwater. Were half the channel to be drag-lined, the hydraulic characters would be changed and waterflow would accelerate, diminishing the channel's intended purpose. The model for the plan predicts that the combination of the detention basins and creek restoration could reduce sedimentation to the lagoon by 45 percent (Figure 3).

Our studies showed that increased urban peak storm flows had dug a 20 foot deep gully in Buena Vista Creek. This massive hole was a primary source of sediment to the lagoon. If no changes were made complete build-out of the watershed would cause further down-cutting and widening of this gully with loss of riparian habitat and rapid filling of the lagoon. The plan recommends the construction of several large stormwater detention basins in the watershed. These basins consist of a small dam or a road crossing with a restricted outlet. The basin allows water to pond during storms and to be slowly released. The computer model showed installation of these basins alone would produce an estimated 20 percent reduction in sediment deposition in the lagoon. Creek Restoration The model for the plan predicts the most cost effective way to lower storm flows to restore those areas of remaining unimproved creek to function as a natural floodplain. The basic idea is to rebuild the creek to a fully vegetated channel with a series of drop structures which create small hydraulic drops and slow water flow. The natural function of the riparian forest in slowing water is used and the acreage of riparian habitat increased throughout the creek corridor. Because portions of the remaining creek channel cross developed

Figure 3 Comparison of a natural watershed where water flows and sediment are in balance and an urban area where impervious surfaces cause increases in water flow, upsetting the balance and resulting in sedimentation of the wetland and loss of riparian forest.

206

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 2 Cross Section of Proposed Enhanced Buena Vista Creek Channel 207

This channel design is revolutionary and downright frightening to many civil engineers used to the standard trapezoidal cement channel. The mathematical calculations inherent in this design are complex and use the current state of the art in hydraulic and hydrologic modelling. But this design will serve as a model for many of the surrounding communities whose waterways and wetlands suffer the same problem. In addition, the local government intends to re-develop its downtown facing onto this green corridor, an encouraging trend. A final recommendation of the plan is to preserve a large riparian forest along the lower creek near the lagoon. Preserving natural floodplain even when it involves expensive acquisition is a more cost-effective sediment control measure than dredging and disposing of spoils from the lagoon. The Conservancy is currently implementing the plan recommendations by having detailed creek designs prepared in conjunction with the city of Vista. Several of the detention basins have been built or are under construction. The local government has cooperated in using this vegetated channel design largely because of the permit requirements of the Army Corps of Engineers and streambed alteration agreement Terms of the California Department of Fish and Game requiring riparian mitigation. The Conservancy intends to maintain highly accurate records for design, construction, and maintenance costs for this channel for comparison with standard hard lined channels. Through the implementation of this plan the life of Buena Vista Lagoon will be extended by at least a hundred years, and Buena Vista Creek will become a restored riparian habitat.

Figure 4 The Lagunitas Creek delta progressively stretches into Tomales Bay at the rate of 50 feet a year. In the last 200 years, the Tomales Bay watershed has undergone a number of significant changes. When the first Europeans arrived, its watershed was largely native grassland with some tracts of redwood forest oak woodland and many streams lined with riparian forest. The thick rooted native perennial bunchgrasses stayed green year round and created a thick mulch over the soil. The Spanish first introduced cattle to Tomales Bay in the early 1800's. The American influx to California brought dairy farms, sheep grazing, and potato farms. By the turn of the century, intensive grazing had practically eliminated the native grasses. Introduced annual grasses from southern Europe, North Africa and Australia, were adapted to heavy grazing and replaced the native prairie. The displacement of native perennial grasses has had a major effect on the hydrology of the watershed. The dense perennial grasses screened the soil from surface erosion and provided a dense fertile mulch which retained moisture. Annual grasses do not provide as much erosion protection; the mulch washes away and the soil layer becomes thinner and less productive. These changes result in greater surface runoff during storms, higher soil erosion rates, and less storage of water in the soil. Higher peak flood flows occur during winter storms eroding channels and under cutting riparian forest. Summer flows in creeks are lower reducing the ability of riparian species to survive and recolonize denuded areas. Through time, the form and structure of the streams feeding Tomales Bay has significantly altered. Historic and continued activities that have contributed to the accelerated erosion of the watershed include: Livestock trampling of streambanks and grazing of riparian seedlings has denuded many streams and caused streambank erosion. Extensive logging began in the 1860's, which exposed the soil to direct erosion and thinning of the soil

Tomales Bay

The Problem

Tomales Bay is one of the largest estuaries in California. Its 223 square-mile watershed is composed of agricultural land, dairies and cattle ranches. However, despite the differences in land use, the Tomales Bay watershed suffers from many of the same problems as that of Buena Vista Lagoon. The two primary tributaries, Lagunitas Creek and Walker Creek, have filled Tomales Bay with tons of sediment and the deltas of these creeks continue to grow at the rate of 35-50 feet per year (Figure 4). The source of this sediment can be traced to historic and present land uses.

208

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

mantle, as well as initiated landslides on steep slopes. Such landslides can remain active for long periods and contribute very high volumes of sediment to stream courses. Extensive potato farming occurred in the Tomales area during the late 1800's. Tilling of the soil for crop cultivation caused a major increase in surface erosion. Poorly located roads often initiate landslides, improperly designed road culverts frequently cause the formation of gullies. Overgrazing loss of native grasses and compaction of soils increase the rate of water runoff resulting in increased peak storm flows. These increased flows have caused severe arroyo formation along Walker Creek with nearly complete loss of riparian habitat. Livestock trampling of streambanks and grazing of riparian seedlings has denuded many streams and caused streambank erosion. If current sedimentation rates continue, the tidal volume of Tomales Bay would be reduced by 10 percent over the coming 50 years. This would be a significant amount in terms of the physical size of the bay.
Erosion Control Plan

reservoir. The Walker Creek watershed also shows highly compacted soils, numerous large gullies and roads with inadequate drainage. Repair of many erosion sites has involved a great deal of riparian revegetation often combined with structural stabilization measures. In many instances water flows or erosion problems are too great to be stabilized with plantings alone; rock or other hard surface stabilization is required to accompany plantings. Repair of large gullies, which can erode tons of sediment in a single storm is a major component of the plans for both creeks. Gully repairs typically consist of rock, concrete or redwood checkdams-small structures which produce a staircase of small hydraulic drops within the gully and significant slowing of water flows. Once the water is slowed and its erosive force lowered, each stair step fills in with sediment. Willow sprigs are planted along side walls and headcuts. Willow can grow in near vertical areas and is easy to establish. The gully is then fenced to exclude livestock so that riparian plants can get a firm hold. Often road repairs accompany gully repair as inadequate road drainage or poorly located or graded roads can cause gullies to form. Riparian vegetation alone can be used to stabilize smaller, less incised gullies. Streambank stabilization and revegetation, is a major component of the program, on Walker Creek the it there it will be a three-step-process. Until the problem of increased peak stormflows is solved, riparian vegetation along the creek will continue to be undercut and wash away. The easiest way to slow channel downcutting may be the installation of large concrete checkdams (grade stabilization structures) which produce a hydraulic drop and slow water as do small checkdams. These checkdams will be located in the upper tributaries. Along the lower portions of the channel rock riprap bank stabilization combined with riparian plantings will be used. Simultaneously, range management plans for ranches in the watershed are being discussed with landowners. The purpose of these plans is to alter the concentration of livestock on already compacted soils through cross-fencing of pasture, development of additional water supplies, fertilization planting of native bunchgrasses and yearround rotation of livestock. If overall vegetative cover can be increased through these measures the stormwater percolation may increase and peak stormwater flows be reduced. Once the erosive problems along the creek are reduced, intensive planting and exclusionary fencing can begin. Establishment of a thick riparian forest will be aided by the year-round flow of water and should greatly help to stabilize steep banks and reduce sediment deposition in Tomales Bay. Since the completion of the Lagunitas Creek restoration project in 1986 our monitoring has compared the

The Conservancy, working in conjunction with the Marin County Resource Conservation District, a local agency which undertakes soil erosion control projects, completed erosion control plans for the two main tributaries to Tomales Bay. As with Buena Vista Lagoon, the expense of removing accumulated sediment from Tomales Bay, far exceeds the cost of controlling the sources of erosion. Consequently the Conservancy has focused on repairing the causes of erosion, a task which involves extensive riparian revegetation. These plans surveyed each subwatershed, mapping erosion problems and their prospective repairs. The Conservancy has granted over $1,200,000 dollars to implement these plans. Lagunitas Creek, which drains the southern portion of the watershed carries large sediment loads. Its water shed includes some agricultural land, state park land and many rural residential developments. Sediment comes from inadequately drained roads, gullies, streambank failures and intensive uses such as horse corrals next to the creek. Walker Creek, which drains the northern portion of the watershed, is almost entirely agricultural land. Erosion occurs in the channel itself; the creek is wide and braided with 15 foot vertical banks and limited riparian forest, despite year-round releases from an upstream

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

209

many different types of repair measures installed. Many of the most successful repairs combine some structural control measure (small checkdam rock riprap) and planting. Deep incised gullies are now forests of willows and Tomales Bay receives less sediment. The repairs on Walker Creek are in their third year and are expected to take another three to four years to complete.

References
Applegate, June et al 1985. Buena Vista Lagoon watershed sediment control plan. Oakland, CA: California State Coastal Conservancy. Chatham, Steven and Liza Prunuske 1986. Walker Creek Watershed erosion control plan - Marin County, CA: Marin County Source Conservation District. Hecht, B. 1980. Substrate Enhancement/Sediment Management Study, Lagunitas Creek, Marin County. Phase II: Sediment Transport and Substrate Conditions 1979-80. Technical report, Marin County CA: Marin Municipal Water District, September 1980. Prunuske, Liza, 1988. Groundwork: A Handbook for Erosion Control in North Coastal California. Marin County Resource Conservation District, Marin County, CA: Storm, D. 1972. Environment of Tomales Bay, Washington D.C. Conservation Foundation. Wahrhaftig, C. 1972 Environment of Tomales Bay, Washington D.C. Conservation Foundation 1972.

Conclusions
The State Coastal Conservancy through our mandate to preserve and enhance coastal wetlands in California has become involved in extensive restoration projects in coastal watersheds. Our studies and experience have found that repairing the sources of erosion is a more cost effective strategy than attempting to dredge sediment from the estuary or wetland. The repairs cited in the two examples involve stabilizing the stream system using riparian forest and limited structural control measures. Watershed management and use of the natural function of riparian forest to slow water and trap sediment is a new strategy in preserving coastal wetlands. As the Conservancy continues to implement these types of projects this strategy will benefit both the riparian and coastal wetland resources.

210

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

RESTORING AND MAINTAINING RIPARIAN HABITAT ON PRIVATE PASTURELAND 1


Nancy Reichard2 Abstract: Protecting riparian habitat from livestock grazing on private land is a complex task that requires paying attention to sociological and economic as well as physical and biological factors. Six livestock exclusion fencing projects on private property in northwestern California are described. The importance of long term maintenance and the need for landowner incentives are discussed. Significant gains may be made via a statewide, coordinated effort to encourage the protection of riparian habitat on private property.

Several dozen coastal streams in Humboldt and Del Norte Counties have reaches that lack riparian vegetation, due at least in part to livestock impact (Streamfellow and Reichard 1983). Many other stream reaches are bordered only by a residual canopy of vegetation. We have observed, as has Shanfield (1984) and others, that riparian stands to which livestock have free access usually lack understory vegetation. Little work has been done to describe the ecology of North Coast riparian systems. In his thesis work on habitat relationships among riparian forest birds in the Eel River Delta, Kelly (1987) provides one of the only detailed analyses. Undisturbed, mature, North Coast riparian vegetation typically is comprised of a dense, diverse understory of herbaceous and woody plants. A canopy of deciduous and/or coniferous trees may include Sitka spruce (Picea sitchensis), redwood (Sequoia sempervirens), black cottonwood (Populus trichocarpa), red alder (Alnus rubra), bigleaf maple (Ater macrophyllum), and willows (Salix spp.) Younger stands have less species diversity. Willows are typical pioneers in disturbed areas. (Kelly 1987, Ray and others 1984, Roberts 1984).

A muddy stream devoid of riparian vegetation and trampled by livestock is a raw wound, in the eyes of those knowledgeable of stream and stream-side ecosystems. The wound can be healed if fences are constructed to exclude livestock from the riparian corridor. The cessation of grazing allows existing or new vegetation to grow. Although there may be several physical and social challenges to creating a fenced riparian corridor on private property (Reichard 1984), once these are overcome, the benefits may be observable within months, due to the resilient nature of some riparian plant species. The Natural Resources Services division of Redwood Community Action Agency (RCAA) constructed six livestock exclusion fencing projects on private property between 1982-1986. The long term success of these projects depends almost entirely conmaintaining an intact fence simple inconcept but complex in practice. This paper summarizes RCAA projects and recommends measures for improving riparian habitat protection.

Projects Implemented by RCAA


The livestock exclusion fencing projects undertaken by RCAA have enclosed a total of approximately 37 hectares of habitat, along six different streams (fig. 1). Approximately 10 hectares are riparian and 5 are instream. The projects were funded by either the State Coastal Conservancy (SCC) or the Department of Fish and Game (DFG). Landowners and the California Conservation Corps (CCC) provided substantial in-kind contributions to several of the projects. At all of the project sites, most of the original riparian vegetation had been cleared at least a generation before the present landowners took charge. Some project characteristics are presented in table 1. Figures 2-5 depict typical preand post- project conditions and will be referred to later.

The North Coast Setting


Because of topography and land use patterns, streams which flow through pastures in Humboldt and Del Norte Counties are almost entirely coastal. In this paper, "pasture" refers to a fenced plot of relatively level land used for grazing livestock, usually relatively intensive due to the limited size of the plot. Most pastureland in this region is used by dairy and beef cattle operations. Residential area streams may be impacted by numerous small pastures used for the "family" horses or cows.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Director of Natural Resources. Redwood Community Action Agency. Eureka, Calif.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

211

reach maps of McDonald and Freshwater Creeks, and a year-long avian census on Tryon Creek. Five of the projects have to date resulted in the exclusion of livestock from the fenced riparian corridor, and in the subsequent establishment and growth of significant numbers of native woody plants. Livestock have not yet been completely excluded from the Little River site. Maintaining these projects is addressed in agreements that RCAA has executed with each of the landowners. Both SCC and DFG require that a license agreement between the landowner and the contractor (RCAA) be recorded on the landowner's property deed. In the agreement, fencing and planted vegetation are defined as "improvements." The landowner agrees to maintain the improvements and not to allow their alteration, for a period of 20 years for an SCC-funded project and 10 years for a DFG project. The agreements are recorded so that they will be effective even if the property changes ownership. In practice, maintaining most of RCAA's projects has been a collaborative effort. The CCC has provided invaluable assistance with maintenance and improvements at the Tryon, Prairie, and McDonald Creek sites. A commercial fishing group has helped to maintain the Little River project. The landowners have played a minor to major role in project maintenance. Their attention to maintenance is proportional to their level of vested interest in the project, and can also vary, dependent on whether or not the landowner lives on the property, leases it, or has a caretaker. Of all of RCAA's projects, Tryon Creek has the most supportive landowner. As a fish and wildlife enthusiast and a third-generation cattle rancher, the landowner had for several years wanted to reestablish vegetation along the stream. Financial support from the SCC made the project possible. Substantial instream habitat improvements were also made as a part of this project. Figure 1 Location of RCAA livestock exclusion fencing projects. All fences were constructed with split redwood posts and five strands of barbed wire. Vegetation was planted at all sites. Willow cuttings from on-site, red alder, and sitka spruce were planted predominantly, along with big leaf maple and redwood. The tree seedlings were obtained from commercial nurseries in northern California and southern Oregon. Project sites are monitored informally by RCAA staff, except at McDonald Creek, where a monitoring project is under way as a part of the SCC contract. Observations at McDonald Creek are being made annually for a 5year period, including stream channel cross sections, vegetation transects, and photographic documentation. Baseline data collected at project sites included stream 212 The Prairie Creek landowner was primarily interested in a project because he was losing pasture to streambank erosion. If he had been able to finance bank stabilization on his own, he probably would have used traditional, non-vegetative measures. He had, in fact, started to install some car bodies along the bank several years ago, but was halted by DFG. He was willing to convert several acres of pasture into a protected riparian corridor in return for our publiclyfunded services used to apply various bio-technical stabilization measures to his banks (figs. 2 & 3). Unlike many landowners RCAA has talked with, he recognized the role that vegetation can play in maintaining streambank stability. As a secondary benefit, he knew that streamside fencing would trap flood-borne woody debris, which he previously had to clear from his pastures after each flood.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 1 RCAA livestock exclusion fencing projects. Stream name Tryon Prairie 1985 SCC Resident 2,650 3 0.8 2.4 3.2 2 yes McDonald 1983 SCC Mgr. 2,440 11 2.8 0.8 3.6 3 yes Little R. Freshwater 1984 SCC Non Res 79 1 0. n/a 0. no 1986 DFG Mang'r 610 8 0.4 0.4 0.8 1 yes Francis 1983 SCC Resident 335 12 0.4 0.2 0.6 1 no Date completed 1986 Funding source SCC Management1. Resident Fence length (m)2 4,100 Set-back(m)3 12 Hectares enclosed:4 Riparian 4.8 Active channel 1.2 Total 6.0 Water accesses6 1 Bank stabilization7 no
1 2

Property managed by, in order: resident landowner, res. manager, non res. landowner. lncludes fence on both sides of stream, where applicable. 3Average, from edge of active channel. 4All acreages are approximations. 5Only one side of the channel was fenced. Stock have access to the stream from the other 6Barriers that allow access to or across the stream but not into the fenced corridor. 7Streambank stabilization measures applied as part of project.

The initial project that RCAA implemented included fencing and planting only. Subsequently, we determined that streambank erosion was going to take out some of the fencing and vegetation before natural healing processes could take place within the corridor. Additional SCC funds and CCC labor enabled us to apply rock riprap and additional vegetation, which has effectively controlled most of the erosion.

Figure 2 Bank stabilization along Prairie Creek. Willow logs and hog wire were staked to what was a bare, eroding streambank. January 1984. Both of these projects allowed the landowners and public-interest groups to work together to solve resource management problems and to learn from each other. Symbolic of these opportunities is the day at the Prairie Creek dairy ranch that a CCC crew of primarily urban kids from southern California, RCAA staff and the landowner planted several thousand trees, saw steelhead spawning, and watched a calf being born in the barn. The McDonald Creek property was being managed for both agricultural and recreational uses. Establishing trees along the stream and improving fish habitat makes the area more attractive to visitors, which was an acceptable trade for the loss of pasture (figs. 4 & 5). Figure 3 Same site as in figure 2, June 1984. Note livestock exclusion fencing installed along top of streambank. As of 1988, willows and alders at this site were six to eight feet tall.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

213

alders and Sitka spruce are thriving. The project was acceptable to the landowner primarily because the exclosed area is not high-quality pasture. A proposal to fence the left bank, although acceptable to the (different) landowner, was met with resistance from the pasture lessee. A barbed wire fence was not acceptable because of the potential for injury to a spirited horse. A pump was not acceptable as an alternative source of drinking water for livestock because it would not be as reliable a water source as the river. An electric fence powered by a solar charger was installed, but did not stand up to woody debris carried into it by streamflow, and dropped on top of it by beavers and overhanging trees. Several shorts were caused by growing vegetation which came into contact with the fence wires. The fence was removed when it became clear that with a sympathetic but absentee landowner, an unsympathetic lessee, and a need for frequent maintenance, it was not going to be effective. As a partial substitute, we designed five "willow pods" oval exclosures averaging 3 meters wide by 12 meters long. The exclosure fence was made with smooth wire and with live willow branches for fence posts. Willow cuttings were planted within the pods. After two years of growth, substantial die-off occurred, possibly due to an inadequate freshwater table and/or to contact with saline water (the site is along the upper part of the Little River estuary.) Cows were able to reach through the smooth wires to browse about two feet within the pods. This reduced the area of effectively protected vegetation noticeably in these relatively small nexclosures. Two pods are still functioning and growing. A commercial fishing organization which operates salmon rearing and habitat restoration projects in the upper watershed planted big leaf maples along the right bank and is planning to construct a strong fence along the left bank in 1988. It is also maintaining contact with the landowner, to encourage the ultimate dedication or sale of the property for conservation purposes, so that livestock exclusion in this challenging location may ultimately be unnecessary (Farro 1988). Figure 5 McDonald Creek, same site as in figure 4, September, 1987. Planted and volunteer red alder is the predominant vegetation. The barbed wire fence constructed along the right bank of Little River does not prevent livestock from crossing the channel from another landowner's property on the left side during low water. Fortunately, because livestock are not always present on the left side and due to steep banks and pools in some areas, stock have not had a complete or continuous impact on the vegetation within the fenced corridor. Some of the planted red The Freshwater Creek landowner was willing to "contribute" pasture for the sake of aiding fisheries restoration. It was probably significant that the property was not a primary source of his income. The Francis Creek landowner agreed to a fencing project as part of a package that included streambank stabilization at an adjacent site.

Figure 4 McDonald Creek, December 1982, just after fencing installed. View is looking up the perennial north fork, mainstem entering from right. Metal gates in the foreground are suspended from a cable, allowing water and flood-borne debris to pass under them, while preventing livestock from entering the protected corridor.

214

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

The Challenges
The physical and biological challenges to restore riparian habitat on private pastureland are minor compared to a broader socioeconomic challenge: securing the stewardship of natural resources on private property when there is no legal or other institutional mechanism for restoring and protecting these resources. In California, the impact of livestock to riparian habitat on private land is virtually unregulated (Sommarstrom 1983). Unless that situation changes, protecting the habitat cannot occur without the voluntary cooperation of the landowners. The total area of privately owned riparian habitat that is impacted by livestock may not be very large; however, along a given stream the damage may be significant. The great number of people involved is both a problem and an opportunity. Landowners with impacted streams can be directly responsible for restoring and protecting the natural resources on their land. Better strategies for encouraging that responsibility need to be developed. Like other resource conservation measures on private land, there has to be incentive and means to implement, and the means and the will to maintain the improvements. Landowners may or may not perceive benefits to be obtained from an exclusionary fencing project. Besides those previously mentioned, benefits of crossfencing and reduced livestock drowning hazard may be provided. Livestock exclusion fencing will always cost the landowner, in reduced pasture acreage. A subsequent cost will be the maintenance of the fence. RCAA has negotiated for fencing projects with at least 12 landowners, all of whom sooner or later decided that the costs outweighed the benefits. In a few instances, persistence has paid off, particularly as landowners become more interested in playing a positive role in the salmon and steelhead trout restoration "movement." Constructing a fence is only the first phase of protecting the streamside corridor. A fence must be inspected regularly and repaired promptly when damaged. Damage from hungry cows in the corridor for just one day can be substantial. (Cross-fencing within the corridor can limit such damage to the segment with the hole in the fence.) Platts (1984a) presents average costs of $60$200 per mile per year for fence maintenance. Using alternative grazing systems to provide riparian habitat protection in lieu of fencing (Platts 1984b) may have applications in some pastureland situations. Implementing and maintaining an alternative system on

private property may be at least as challenging as a fencing project.

Recommendations
Protecting riparian habitat on private pastureland presents unique opportunities and challenges. A statewide, coordinated effort to identify and promote means to restore and maintain this habitat could be very effective, especially if it included landowner representatives. Inventorying pastureland riparian habitat, identifying the potential for restoration, and assessing the costs and benefits of doing so, would help determine just how aggressively this component of riparian habitat restoration should be pursued and what the priorities should be. Because of the nature of the problem, negative incentives i.e. regulations may not be very effective or manageable. As Sommarstrom (1984) put it, mandating fencing might be like trying to legislate morality. On the other hand, there is room for progress in developing and using positive incentives, ranging from education to tax benefits. A compendium of information regarding conservation easements and tax considerations related to livestock exclusion would be useful for both landowners and promoters. Educational campaigns through agricultural, equestrian, and other organizations could make "riparian habitat" a household concept among riparian landowners.

Conclusions
Management of riparian habitat on private pastureland is a broadly interdisciplinary topic. Its complexity is both a blessing and a curse solutions to the problems of protection are not simple, but they hold the promise of improved public/private cooperation. This area of riparian habitat management warrants special and focused consideration by those interested in seeing that the habitat is restored and protected.

References
Farro, Mitch. Little River Project Manager, Pacific Coast Federation of Fishermen's Association/Trinidad Fishermen's Marketing Association, Trinidad. Personal communication. 16 June 1988. Kelly, John P. 1987. Habitat relationships among riparian forest birds in the Eel River delta, California. Arcata: Humboldt State University; 134 p. master's thesis.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

215

Platts, William S.; and Wagstaff, Fred J. 1984. Fencing to control livestock grazing on riparian habitats along streams: is it a viable alternative? North American Journal of Fisheries Management 4:266-272. Platts, William S. 1984. Riparian System/Livestock Grazing interaction research in the intermountain west. In: Warner, Richard E., and Hendrix, Kathleen M., editors. California Riparian Systems. Berkeley: University of California Press. 424-429. Ray, Dan; Woodroof, Wayne; and Roberts, R. Chad. 1984. Management of riparian vegetation in the northcoast region of California's coastal zone. In: Warner, Richard E., and Hendrix, Kathleen M., editors. California Riparian Systems. Berkeley: University of California Press. 660672. Reichard, Nancy. 1984. Riparian Habitat Restoration: Some techniques for dealing with landowners, livestock, and eroding streambanks. In: Hassler, Thomas J., editor. Proceedings: Pacific Northwest Stream Habitat Management Workshop; 1984 October 10-12; Arcata, CA. Available from: Calif. Cooperative Fisheries Unit,

Arcata, CA. 128-137. Roberts, R. Chad 1984. The transitional nature of northwestern California riparian systems. In: Warner, Richard E., and Hendrix, Kathleen M., editors. California Riparian Systems. Berkeley: University of California Press. 85-91. Shanfield, Allan N. 1984. Alder, cottonwood, and sycamore distribution and regeneration along the Nacimiento River, California. In: Warner, Richard E., and Hendrix, Kathleen M., editors. California Riparian Systems. Berkeley: University of California Press. 196-209. Sommarstrom, Sari. 1984. Riparian regulations: random, redundant, or rational? In: Warner, Richard E., and Hendrix, Kathleen M., editors. California Riparian Systems. Berkeley: University of California Press. 275-280. Streamfellow, Dwight; and Nancy Reichard. 1983. HumboldtDel Norte coastal streams restoration project, final report, prepared for the California State Coastal Conservancy. Redwood Community Action Agency, Eureka, California. 30 p.

216

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

RECOVERY OF RIPARIAN VEGETATION ON AN INTERMITTENT STREAM FOLLOWING REMOVAL OF CATTLE 1


Jerry J. Smith 2 Abstract: In 1984 1987 the recovery of riparian willows (Salix spp.) and sycamores (Platanus racemosa) was studied on two short, intermittent stream sections in a newly acquired portion of Henry W. Coe State Park in central California. Prior to removal of cattle in 1983, the plots contained mature sycamores, one young sycamore, and five willows. By 1985 over 320 willows, 16 sycamores and 1 cottonwood (Populus fremontii) had appeared, and basal sprouts had developed on the mature sycamores. Young willows and sycamores grew slowly, and establishment and growth generally ceased as surface flows disappeared. Because of slow growth at the sites, a significant willow corridor is probably only possible in the absence of cattle browsing. of and in the rain shadow of both the coast range and the highest ridges of the inner coast range. Vegetation within the 200 to 800 m high watershed is primarily: chaparral, dominated by chamise (Adenosloma fasciculatum); oak woodland, dominated by blue oak (Quercus Douglasii), buckeye (Aesculus californica), California bay (Umbellaria californica), and digger pine (Pinus sabiniana); and grassland, dominated by introduced annual grasses (Avena spp. and Bromus spp). The stream is intermittent and generally dries to a few isolated pools by early to mid summer. During the study surface flow ceased by late June of 1984, May of 1985, late May of 1986, and March of 1987. At the start of the study, woody riparian vegetation consisted almost exclusively of mule fat (Baccharis viminea) and mature sycamores (greater than 25 cm in diameter).

Grazing has been found to have severe effects upon riparian habitats (Meehan and Platts 1978). Altered species composition and vegetation density, along with streambank erosion and channel widening, have stimulated research on fencing (Platts and Wagstaff 1984) and seasonal regulation of grazing (Bryant 1985; Siekert and others 1985) as ways to reduce impacts. However, most studies of grazing impacts and treatment strategies have been conducted on public lands outside of California. A century or more of intensive private ranching has probably substantially changed parts of the California landscape, but left us with few undisturbed areas with which to judge the impacts of cattle. In 1982 the California Department of Parks and Recreation acquired two former ranches as additions to Henry W. Coe State Park. Cattle grazing was halted in 1983, but a heated public debate over the general plan for the park included consideration of the desirability of renewed grazing. The absence of documented evidence for grazing damage in central California was used by some as proof that no damage occurs. The removal of cattle in 1983, however, provided an opportunity to assess the effects of longterm cattle grazing.

Methods
In March 1984 all young willows greater than 5 cm tall in a 35 m section of North Fork Pacheco Creek (study area A) were tagged with numbered Floy tags, and heights of longest stem recorded. The plot was resurveyed again in July and October 1984, May and October 1985, and October 1987, and heights of tagged and untagged willows recorded. New willows taller than 30 cm and a portion of shorter willows were tagged during resurveys, and evidence of browsing and general plant health (dead branches, desiccated leaves, etc.) were recorded during each survey. In May 1985 study area B, a 300 m long stream section 0.1 km upstream of study area A, was established, and all willows, cottonwoods, and young sycamores were tagged with numbered Floy tags, and heights and health recorded. Study area B was resurveyed in October of 1985 and 1987. Five willows and one young sycamore were present in the streambed between the two study areas prior to removal of cattle in 1983. These 6 young trees plus 20 associated mature sycamores were cored with an increment borer to determine approximate ages of the trees. Ages determined from cores were considered approximate for the young trees, as subsequent studies suggested that several years of suppressed growth, with possibly no growth rings, could have occurred during early years due to cattle browsing. Ages of mature

Study Area
North Fork Pacheco Creek is a fourth order stream in the eastern portion of Henry W. Coe State Park in the Mt. Hamilton Range of central California. Although a tributary to the coastal Pajaro River, the stream is east
1Presented 2

at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

Lecturer in Biology, San Jose State University, San Jose, California

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

217

sycamores were approximated by extrapolating growth rates found in the outer portion of the core to the rotted interior found in all trees. Cores were coated with phloroglucinol to improve their readability. Numbers of basal sprouts on mature sycamores were counted on 50 trees in October 1984 and October 1985 for comparison with 1983, when sprouts were absent. Sprouts were not counted in October 1987, but general observations were made on condition of sprouts within the two study areas.

Table 2 Heights (cm) and numbers of willows and young


sycamores in study area B on North Fork Pacheco Creek. Willows 1985 Height 5 - 49 50 - 99 100-149 150-199 200-299 300-399 400-499 500-599 Totals: May 3 12 5 8 13 4 Oct 10 12 3 11 9 6 1 45 52 1987 Oct 5 13 6 4 8 8 1 45 15 16 16 May 1 2 2 4 5 1 Sycamores 1985 Oct 2 2 2 4 3 2 1 1987 Oct 1 2 2 8 1 2

Results
Photographs of study area A taken in April 1983 showed no plants larger than 30 cm, other than mulefat and mature sycamores, but by March 1984, 70 willows to 130 cm high were found (table 1). By July of 1984, 271 willows were present, with 3 exceeding 150 cm tall. No new willows appeared by October and two tagged willows had died. In 1985, no new willows appeared and 26 died. By October 1985, nine willows exceeded 150 cm in height and the tallest was 265 cm high. By October 1987, only 86 willows remained, with most small willows apparently buried or washed away by an intense February 1986 flood. Fifteen willows died in 1987 due to desiccation. Photographs of study area B in April 1983, showed no plants larger than 30 to 50 cm tall. By May 1985, 45 willows, with two-thirds taller than 1 m, were present (table 2). Fifteen young sycamores and one cottonwood were also present. By October of 1985, seven additional willows and one additional sycamore were present. Two new sycamores appeared in 1987, but Seven willows and two sycamores were apparently lost to 1986 flooding. Study area B had a lower density of young trees than study area A, but the trees grew faster. By October 1987, the majority of willows and all but 1 sycamore exceeded 1 m in height, and 12 trees exceeded 3 m in height (table 2). Table 1 Heights (cm) and numbers of willows in study area A
on North Fork Pacheco Creek. Date 1984 Height 5-9 10 - 19 20 - 29 30 - 49 50 - 99 100-149 150-199 200-299 300-399 Totals: Mar 7 35 6 13 8 1 Jul 73 63 40 46 36 10 2 1 Oct 65 67 43 45 36 10 2 1 1985 May 45 53 37 67 38 11 3 1 Oct 10 38 51 74 49 12 5 4 1987 Oct 7 8 5 23 23 13 4 2 1 86

Most willows tagged in March of 1984 showed substantial growth by July, despite limited browsing by rabbits and deer (fig. 1). Surface flow in study area A was gone by late June of 1984, and few willows showed growth between July and October (fig. 2). All willows showed some browsing by October, and some larger willows suffered limited damage from sparring male deer. In 1985 surface flow ceased by May in both study areas, and only larger willows and sycamores showed May to October growth (fig. 3).

70

271

269

255

243

Figure 1- Growth of browsed (triangles) and unbrowsed (circles) willows from March to July 1984 in study area A.

218

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

growth from October 1985 to October 1987, with 5 of 14 showing height gains of 50 percent or more (fig. 4). The sycamore showing no height gain was broken off near the base, apparently during the February 1986 flood. Three of the five willows present before cattle removal were dated by cores to 1977. The other two were dated to 1978. The single young sycamore present prior to cattle removal was dated to 1977. Estimated ages of mature sycamores ranged from 79 to 135 years.

Figure 2 Growth of willows from July to October 1984 in study area A. From October 1985 to October 1987 willow recovery was hindered by browsing by trespass cattle. Although few cattle were present, they spent much of their time along the stream. The problem was especially pronounced in 1987, when nine cattle were present on and near the study areas. Although 1986 was a relatively wet year, growth of willows from 1985 to 1987 was generally poor, especially on study area B, where the cattle spent more of their time (fig. 4). In study area B, only half (21 of 42) of tagged willows showed any growth from October 1985 to October 1987, while nine willows were reduced 20 to 70 percent in height by severe cattle browsing. Willows 1 to 2 m high in 1985 suffered most, with 8 of 12 being reduced in height and 4 being reduced more than 40 percent. In study area A, heavy 1985 1987 tag loss limited height comparisons, but all 4 willows which were 1 m to 2 m high in October 1985 and retained tags were reduced in height by October 1987 (5 to 45 percent). Although all willows taller than 3 m showed some growth from 1985 to 1987, most, including the five large willows present prior to cattle removal, showed pronounced browse lines in 1987. On the 50 sycamores examined for basal sprouts only a total of 46 sprouts were found in October 1984, but by October 1985 the total had reached 197. Sprouts were present on all but 2 trees by 1985. Although trespass cattle heavily browsed willows, there was no evidence that they browsed sycamore basal sprouts or young sycamores in 1987. All but one young sycamore showed

Figure 3 Growth of willows (open circles), sycamores (closed circles), and cottonwoods (square) from May to October 1985 in study area B.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

219

Figure 4 Growth of willows (open circles), sycamores (closed circles), and cottonwoods (square) from October 1985 to October 1987.

Discussion
Prior to removal of cattle, the study sections on North Fork Pacheco Creek contained primarily sycamores germinated prior to 1910 and unpalatable mule fat. The single young sycamore present and three of the five willows date to approximately 1977, and the other two willows date to 1978. It appears that successful survival of young willows and sycamores on this intermittent stream has been limited to a single period within the last 70 or more years of intensive cattle use. The isolated survival of a sycamore and possibly three or more willows in 1977, the second year of a severe drought, is a surprising pattern for riparian species. However, poor forage conditions and surface water availability during the drought resulted in removal of cattle from ranches in the area in 1977 (Peter Andresen, pers. comm.), and 1978 was a wet year, apparently allowing successful establishment. Most of the new willows and sycamores appeared in 1984, and total numbers actually declined from 1984 to 1987 (table 1). The loss of established plants was due to the 1986 flood and to desiccation. The lack of significant recruitment in 1985 and 1987 was apparently the result of wide year to year variation in water availability. Successful willow and sycamore germination sites were mostly at the tails of pools and along riffles, sites where seeds are likely to lodge, and where subsurface water

remains as flows decline. In 1984 surface flows lasted into late June. In 1985 1987 surface flows ceased by May, providing less time for successful seedling establishment. Desiccation losses of established willows in study area A were also much greater in 1985 (26 deaths) and 1987 (15 deaths) than in 1984 (2 deaths). The presence of 131 willows, 16 young sycamores, and 1 cottonwood three years after the removal of cattle indicates that, despite the intermittent character of the stream, portions of North Fork Pacheco Creek have the potential to develop significant riparian vegetation in the absence of cattle. Growth rates, however, were low, and germination and growth of small trees was generally limited to the relatively brief period of surface flow. Because of slow growth and infrequent germination success, willows were unable to establish or survive previous intensive cattle browsing. Even light or seasonal cattle presence might be sufficient to prevent establishment of a significant willow population; in 1985 1987 many established willows suffered severe hedging by only a limited number of trespass cattle. Young sycamores and basal sprouts on mature sycamores are apparently not preferred food and were not browsed by the few trespass cattle present in 1985 1987. However, young sycamores were absent and basal sprouts very rare or absent during the period of intensive grazing prior to 1984. The sycamore population might be able to maintain itself under a regime of light cattle stocking or in a wetter area, where young sycamores grow faster. However, previous persistant heavy browsing by cattle apparently eliminated recruitment on this intermittent stream. Although sycamores are presently the "typical" riparian tree of many intermittent streams in central California, widespread and intensive cattle grazing may be jeopardizing recruitment. Longevity of sycamores has not yet been studied, but all 20 of the 79 to 135 year old sycamores cored on North Fork Pacheco Creek had partially rotted trunks. For many grazed intermittent streams in California, lack of recruitment and loss of older sycamores makes barren streamsides a real possibility in the future.

Acknowledgements
I thank Colleen Pelles and Mark Robinson, Department of Biology, San Jose State University; and Thomas Taylor, California Department of Parks and Recreation, for assisting with field work. I thank Harry Batlin, California Department of Parks and Recreation, for permission to conduct the study.

References
Andresen, Peter (Personal Communication).

220

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Bryant, Larry D. 1985. Livestock management in the riparian ecosystem. In: Proceedings of the North American Riparian Conference; April 16-18; Tucson: Univ. Arizona. Meehan, W. P.; Platts, W. S. 1978. Livestock grazing and the aquatic environment. J. Soil and Water Cons. 33(6): 274-278. Platts W. P.; Wagstaff, F. J. 1984. Fencing to control

livestock grazing on riparian habitats along streams: is it a viable alternative? North American J. Fisheries Management 4: 266-272.

Siekert, Ronald E.; Skinner, Q. D.; Smith, M. A.; Dodd, J. L.; Rodgers, J. D. 1985. In: Proceedings of North American Riparian Conference; April 16-18; Tuscon: Univ. Arizona.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

221

GIANT REED (ARUNDO DONAX): A CLIMAX COMMUNITY OF THE RIPARIAN ZONE 1


John P. Rieger and D. Ann Kreager2 Abstract: Active management of coastal streams is needed to ensure the continued existence of significant riparian systems in Southern California. The concept of a dynamic self-replacing plant community is no longer a truism there. In the past decades one exotic species in particular, the Giant Reed (Arundo donax L.) has had an ever-increasing negative role in the succession of riparian systems. The aggressiveness of this exotic has enabled it to invade disturbed areas along many watercourses of Southern California. Giant Reed is also capable of invading mature woodlands, interrupting the cycle of regeneration normally experienced in river systems. Giant Reed stands can become climax communities, replacing natural riparian habitats. Without active management of the vegetation, the survival of many riparian residents, including some endangered species, may be at risk. Mitigation by replacement of lost habitat must be combined with proper management of the areas surrounding those sites lest Giant Reed communities claim much new acreage.

Methods
Three locations on the San Luis Rey River and the San Diego River in San Diego County were studied. Sites within these locations were chosen for sampling purposes to include riparian vegetation of different known ages and habitat quality. Frequency and percent cover of Arundo were determined on a stretch of river immediately east of the State Route 76 bridge crossing on the San Luis Rey River. A 610-meter transect was established parallel to the river's edge. Some 40 7.6-meter diameter quadrats were located at random distances away from the river off this transect at 15.2-meter intervals. Measurements were made within the quadrats to estimate frequency, percent cover, and average maximum height of Arundo. Colony distribution was determined using an aerial photograph (1:600 scale) covering an area approximately 6.1 hectares on the San Luis Rey River. A belt transect 365.9 meters by 15.2 meters (0.56 hectares) was randomly located in each of four equal cells oriented parallel to the river's edge. A digital planimeter calculated the area comprised of Arundo. Plant maximum heights were measured at several sites on both rivers using a telescoping metric measure rod. Individual stands of Giant Reed were randomly selected and the maximum heights of each stand recorded and averaged. Maximum average heights were not obtained in riparian habitat where A. donax colonies were established 4 years or less. Average growth rates were determined by measuring plant heights at various intervals after existing Arundo stands were cut back to soil level. Ages of Arundo stands were determined from aerial photograph series and personal knowledge of the authors. More detailed vegetation data were collected on the San Diego River east of Mast Street Bridge adjacent to Mission Trails Regional Park in Santee. Measurements of randomly located circular quadrats and line intercept analysis were used to estimate density, dominance, frequency, and cover values for all species encountered.

The floristic structure and composition of riparian forests can provide examples of developmental patterns and successional stages within a biotic community that may be indicative of the system's future. Because vegetation dynamics of a riparian system remain in a state of perpetual succession in part due to the reshaping of the riverbed by storm events and less frequent high flood events, it may follow that plant succession is defined by the presence or absence of individual species. Disruption of vegetational succession by social, commercial, and agricultural development can alter the physical and biotic composition of a riparian system to the extent that natural regeneration often cannot occur, thus encouraging the proliferation of exotic species. Arundo donax, or Giant Reed, was introduced to the California landscape in the 1800's. Due to the prolific nature of this grass, presence of this species within a riparian community is becoming increasingly common. Examination of the properties of this invasive plant suggests that proper management of a riparian system which contains Arundo will involve active participation of control and eradication to maintain a heterogeneous plant community.

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. District Biologist, California Department of Transportation, San Diego, Calif.; Ecological Consultant, California Department of Transportation, San Diego, Calif. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

222

Results
A. donax growth appears quite rapid at the onset of colony establishment, especially if growth occurs from previously established rhizomes. The average maximum stem height of observed A. donax stands of various ages expressed in mean height and standard deviation were as follows:
Mean3 40 days1 150 days1 4 years2 35 years2
1

Table 2 - Circular quadrat vegetation analysis on the San Diego River.


Dom. Vegetation Group Native Trees Native Shrubs and Herbs Arundo donax Exotic Shrubs and Herbs 0.3 0.02 0.1 1.5 20.0 0.01 1.4 --6.8 2.8 0.9 0.1 1.3 --92.3 6.4 57.0 11.9 26.0 6.4 Rel. Dom. (%) Freq. Rel. Freq. (%) % Cover

SD 0.28 0.74 1.12 1.27

2.5 4.0 3.8 5.9

Duration of growth since cutting of established 2 Original stand. 3 Height in meters.

A maximum average height of approximately 6 meters was reached in the mature riparian zone which corresponds to maximum heights reported by the University of California Agricultural Extension Service (Fischer 1983). Growth rates of A. donax from established rhizomes averaged 6.25 cm/day (S.D 0.7) for 40 days' growth. And 150 days of growth averaged 2.67 cm/day (S.D. 0.49). The area occupied by Arundo donax and its pattern of distribution along a section of the San Luis Rey River is presented in table 1. Although the area occupied by Giant Reed throughout this stretch of established riparian habitat is relatively small, a dense distribution occurs in the belt transect closest to the riverbank. The belt transect closest to the river incorporated 97.1 percent of the Arundo encountered. Giant Reed occupied 6.0 percent of all transects studied. Quadrats sampled further from the river show a dramatic decrease in presence of Arundo.

If dominance is defined as occupation of the largest basal area per area sampled it is evident that A. donax within this riparian zone is dominant only in the vicinity adjacent to the riverbank. On the San Diego River, native flora was dominant and provided the greatest amount of cover of the 4 vegetation groups used to classify the riparian community. However A. donax was encountered far more frequently than any other vegetative group. Relative frequency of Giant Reed on the San Luis Rey was approximately 30 percent less than its relative frequency on the San Diego River and provided 10 percent less cover (table 2). Arundo donax comprises a significant proportion of the riparian habitat on the San Luis Rey and San Diego rivers of San Diego County. Regardless of quadrat distribution, more than half of the quadrats sampled along both drainages contained Arundo. Both areas sampled were within established riparian zones however the amount of disturbance and degradation varied between drainages.

Discussion
A. donax does not invest large amounts of energy in the development of an extensive woody root or branch system which may account for its rapid rate of growth. Growth rates of 1.02 to 1.52 cm/day were recorded for Salix goodingii and Salix laevigata on the Kern and Lower Colorado River in California (B. Anderson, telephone communication). Recorded growth rates for A. donax are 2.1 to 4.9 times faster. The physical presence of Arundo can inhibit to some degree the establishment or growth rate of native and exotic species often resulting in pure stands of Giant Reed. This situation can be observed readily along all the major river drainages in San Diego county. The fast growth rate and ability to attain heights of between 2.5 and 4.0 meters in less than a complete growing season assures a competitive advantage over slower growing native species. By comparison, willow trees on the Sweetwater River in San Diego county were reported

Table 1 Distribution and Area of A. donax in transects parallel to San Luis Rey River.
Transect Distance from river (meters) Area comprised of Arundo (sq. meters) 1 7.3 2 39.1 3 80.3 4 114.5

1365.6

33.5

7.3

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

223

to have grown 1.5 to 1.8 meters in one growing season (Rieger 1988). Flowering in Giant Reed occurs after intervals of several years, thus propagation is primarily vegetative. Open colonies or groves are established as shoot-producing rhizomes are spread extensively underground (Bailey 1976). New growth can also be established by simple division of the colony; as older stems fall to the earth or are torn from the ground, stem segments can reestablish themselves, producing a new colony. Within a river system the distribution of establishment does not appear to be random (table 1). The highest concentration of colonies occur closest to the river. Frequency and magnitude of the river flow is most likely the major contributing factor influencing this pattern of distribution. Strahan (1983) found that distribution and development of riparian vegetation is regulated by erosion, deposition, and lateral channel migration. River currents create a constant process of erosion with deposition of eroded material occurring further downstream. Flooding, scouring and debris sedimentation serve to promote expansion of A. donax colonies along this zone of frequent inundation. Disruption of a stream system by natural events (flooding) is not the only avenue available to Giant Reed for its establishment into a system. Disturbance from earth-moving activity can encourage the spread of A. donax (fig. 1) even in areas far removed from the water table. Where earth-moving equipment is used the colony can actually spread at a much faster rate than by rhizome growth alone. Evidence can be found in newly graded restoration projects or other construction sites (Rieger 1988). For years Camp Pendleton had a program of clearing the riparian habitat as part of a water conservation program. Giant Reed present within the community was distributed throughout the area which had been denuded. Expansion of the existent population occurred when this practice had been eliminated thus increasing the area previously occupied by the species (Rieger, pers. obs.). Arundo donax occupies a substantial portion of the riparian system. It far exceeds all other exotic species on the San Diego River and it occurs with a very high frequency on both drainages studied. This high frequency indicates a greater presence within the vegetation community than cover alone would imply. This suggests that the invasive ability of Giant Reed is high regardless of the nature of the existing habitat, though establishment of A. donax is probably limited in more dense and mature riparian stands (fig. 2). Although a few bird species have been observed utilizing the plant for nesting. purposes (Kreager, pers. obs.), the presence of Giant Reed essentially creates a zone devoid of wildlife. The dry climate of San Diego precludes extensive decomposition of its vegetation, therefore even dead, arundo remains and continues to pre-

clude encroachment by plant species with wildlife values.

Figure 1 In a newly graded site along the First San Diego River Improvement Project Arundo donax has become established before native species. Another exotic, the Castor-bean (ricinus communis) can be seen in the foreground.

Figure 2 Arundo colonies at the edges of riparian habitats on the San Diego River.

224

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Several significant differences between A. donax and other perennial exotics exist. Very few form expanding contiguous colonies or attain large heights as do several native species. The dependence of Arundo upon flooding and vegetative propagules has acted as a limiting factor for its invasion into a habitat. The low rainfall and infrequent flooding in Southern California has kept arundo dispersal rates down, however the slowness of dispersal compared to other exotic species is compensated by its permanence following colonization.

principles must be applied for the management and control of A. donax as well if the many endangered and other sensitive species dependent upon riparian communities are to proliferate.

Acknowledgments
We thank Kathryn J. Baird for her assistance in data collection and editorial comments; and John A. Beezley for his editorial comments.

Conclusions
Arundo donax is an extremely fast growing perennial plant reaching mature heights within the first growing season. Dispersal is primarily by vegetative means of stem and rhizome fragments. Flooding is considered the primary natural mechanism of dispersal. Grading and other construction activity can greatly increase the area occupied by Arundo colonies. Restoration sites are easily and quickly invaded especially if the plant was present on site prior to preparation. Arundo is found throughout the river drainages studied. This distribution is viewed as being potentially disastrous for the overall habitat quality of the riparian system. Recommendations Loss of riparian habitats has prompted increased interest in them. The presence of Arundo and its ability to compete successfully in riparian systems indicate a future decline in the habitat quality of riparian systems. The end result may be riparian habitat comprised of large percentages of A. donax with greatly reduced habitat quality. Several California endangered bird species are dependent upon riparian habitat, therefore it is extremely important that active Giant Reed management be implemented. Successful techniques have been developed for eradication of Saltcedar (Tamarix spp.)(Kerpez and Smith, 1987) from significant areas in the desert. Eradication

References
Anderson, B. 1988. Telephone communication. Bailey, L. H. 1976. Hortus Third: A concise dictionary of Plants cultivated in the U.S. and Canada. Staff of the L. H. Bailey Hortorium, Cornell University. New York: MacMillan Publishing Co.; pp. 135-136. Fischer, B. B., Lange, A. H.; Crampton, B., 1983. Giant Reed-Arundo donax L. Grass Family. In Growers weed identification handbook; Oakland, Calif., Univ. of Calif. Agriculture Extension Service. Kerpez, T. A. and N. S. Smith. 1987. Saltcedar control for wildlife habitat improvement in the Southwestern United States. U. S. Dep. of Int. Fish and Wildlife Service. Resource Publication 169. Washington D.C. Kreager, D. A. 1987. Personal observation, author. Rieger, J.P. 1987. Personal observation, author. Rieger, John P. 1988. California Department of Transportation. Riparian restoration projects in San Diego County. In: Rieger, J.P.; B.K. Williams, eds. Proceedings of the second native plant revegetation symposium. [San Diego, Calif., April 15 -18, 1987.] NPRS, San Diego, Calif.,; pp. 213-220. Strahan, J. 1983. Regeneration of riparian forests Central Valley. In: Warner, R. E., Hendrix, (ed.). California Riparian Systems. [Univ. of Cal., September 17-19, 1981.] University of California Berkeley. of the K. M. Davis, Press,

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

225

TECHNIQUES FOR MINIMIZING AND MONITORING THE IMPACT OF PIPELINE CONSTRUCTION ON COASTAL STREAMS 1
Thomas W. Mulroy, John R. Storrer, Vincent J. Semonsen, and Michael L. Dungan 2

Abstract: This paper describes specific measures recently employed for protection of riparian resources during construction of an oil and gas pipeline that crossed coastal reaches of 23 perennial and intermittent streams between Point Conception and Gaviota in Santa Barbara County, California. Flumes were constructed to maintain stream flow; anchored straw bales and silt fences were used to filter sediment in the streams; water bars, incorporated straw mulch, anchored jute fabric, and trench plugs were installed to minimize soil loss on slopes; jute and synthetic fabrics were used to stabilize stream banks; construction-related removal of vegetation was minimized through site-specific project modifications agreed to prior to construction; and impacts on sensitive plant species were reduced by transplanting or by salvaging and replanting after construction. Construction at stream crossings was limited by permit to the summer period of normally low streamflows and was required to avoid the breeding season of sensitive riparian-dependent bird species. Restoration of native woody vegetation was encouraged through a number of approaches emphasizing use of material native to the individual watersheds. Many of the more successful approaches employed in this project resulted from refinements to original plans made in the field by representatives of the applicant, construction contractors, consultants to the applicant, environmental monitors, and representatives of Santa Barbara County.

struction activities were confined to a 100-foot maximum right-of-way (ROW). The construction corridor was narrowed to less than 100 feet (from 20 to 80 feet) at stream crossings where feasible. From a landfall about 1.5 miles north of Point Conception, the pipeline crosses coastal terraces and low hills vegetated primarily by annual grassland dominated by a variety of introduced grasses. Coastal sage scrub, dominated by coastal sagebrush (Artemisia californica) or purple sage (Salvia leucophylla), occurs locally, especially on hillsides leading into drainages. The drainages are typically well-incised into the terraces, with the streambeds generally lying 40 feet or more below the terrace elevation. The stream corridors typically are vegetated by riparian forests or woodlands. These are dominated exclusively by arroyo willow (Salix lasiolepis) along the drier watercourses. In addition to arroyo willow, larger streams may support occasional red willow (Salix laevigata), black cottonwood (Populus trichocarpa), western sycamore (Platanus racemosa), and box elder (Ater negundo subsp. californicum). White alder (Alnus rhombifolia) occurs in the vicinity of the ROW and is confined to areas of perennial flow. Coast live oak (Quercus agrifolia), toyon (Heteromeles arbutifolia) and elderberry (Sambucus mexicana) commonly occur in and adjacent to the riparian zone, but are not exclusively riparian species in this region. Steep hillsides, leading into the streambeds, are occasionally dominated by nearly pure stands of holly-leaf cherry (Prunus ilicifolia), that may be relicts of an earlier occurrence of chaparral. From west to east, the pipeline corridor crosses the Bixby Ranch, the Hollister Ranch, Gaviota State Park and smaller private landholdings near its eastern terminus. Livestock grazing is the predominant land use along the pipeline ROW, except in Gaviota State Park. The Hollister Ranch is subdivided into 100-acre parcels, most of which contain residences. The entire pipeline alignment lies in the Coastal Zone and most of the riparian habitats it crosses are designated environmentally sensitive habitat areas (ESHA) in the Santa Barbara County local coastal plan (LCP) (Santa Barbara County 1982). Among other things, the ESHA designation recognizes the importance of these

The Point Arguello pipeline was recently constructed by Chevron U.S.A. Inc. to convey outer-continentalshelf (OCS) oil and gas from offshore platforms near Point Conception to a processing facility at Gaviota in Santa Barbara County, California. Construction of the onshore portion of the pipeline, which parallels the coastline, involved crossing 23 perennial and intermittent streams along its 16-mile onshore alignment (Figure 1). Distance from the ocean along the alignment ranges from about 1 mile on the west to a few hundred feet at Canada Agua Caliente and Canada de Alegria. The project involved burial of a 24-inch diameter oil pipeline and a 20-inch gas pipeline within a single trench. Burial depth was a minimum of 36 inches to top of pipe, with greater depths required at certain stream crossings in order to protect the pipes from scour during floods. Con-

1 2

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Principal Scientist, URS Consultants, Santa Barbara, California; and Senior Ecologist, URS Consultants, Santa Barbara, California, respectively. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

226

Figure 1 Project location near Point Conception corridors to the overall landscape. They provide food, escape cover, water sources, and movement corridors for a variety of wildlife that use the surrounding upland habitats. A number of sensitive aquatic and avian species were located in these habitats, especially in the larger, less disturbed streams. Sensitive aquatic species occurring in streams crossed by the pipeline alignment

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

227

included tidewater goby (Eucyclogobius newberryi), California red-legged frog (Rana aurora draytonii), and southwestern pond turtle (Clemmys marmoraia pallida), and reported runs of steelhead trout (Salmo gairdneri). The tidewater goby, red-legged frog and southwestern pond turtle are category 2 candidates for federal listing as threatened or endangered. (A category 2 candidate species is one that U. S. Fish and Wildlife Service [USFWS] believes may be appropriate for listing but for which additional information is required.) Preservation, maintenance and improvement of steelhead trout habitat is of high management interest to both California Department of Fish and Game (CDFG) and USFWS. Site-specific information on the stream crossings as well as additional general information on the project region is provided in Mulroy and others (1984).

Maintenance of Stream Flow

Permit conditions (Santa Barbara County 1984) required that construction activity be confined to low-flow periods, basically the summer and early fall months, and that the existing flow be maintained across the construction area at all times. Construction during the low- flow period helps minimize impacts on water quality; maintenance of low flows was judged essential to maintain downstream populations of aquatic biota, including several sensitive vertebrate species. Prior to construction at a given stream crossing, sediment traps were installed as described in the next section, and vegetation was removed from the construction right-of-way (ROW). The initial work involved crews using chain saws, who were able to minimize disturbance to vegetation outside the ROW (e.g., by directional felling of trees). The ROW was generally perpendicular to the stream course and a flume was installed to convey water across the ROW. The flume consisted of a large pipe along the bottom of the streambed extending from a short distance upstream of the ROW to a short distance downstream from the ROW. Project specifications called for construction of a temporary dam, consisting of straw bales and plastic, at the upstream end of the ROW to direct water into the pipe. Generally, it was necessary to have the pipe flush with the stream bottom and to have the opening surrounded by sandbags to minimize leakage. At the downstream end, simple energy dissipation structures were installed to prevent erosion caused by the concentrated flow exiting the pipe. Following sediment trap installation, vegetation removal, and flume installation, pipeline trenching was conducted with backhoes under the flume, which stayed in place for the duration of construction activities at a given stream crossing.
Sediment Control

Environmental Protection Measures


Specific environmental protection measures were developed through the environmental review process and incorporated into the project. These include measures to maintain stream flow, to control sediment, to stabilize stream banks, to minimize construction-related removal of vegetation, to preserve rare plants, to protect sensitive wildlife and aquatic species, and to restore the native vegetation. Some of these measures were originally developed by the applicant and outlined in the original project description; others were recommended by the EIS/EIR preparers. The EIS/EIR review was directed by a joint review panel of permitting agencies, led by the U.S. Minerals Management Service and the County of Santa Barbara. Many of the mitigation measures were originally described in a restoration, erosion control, and revegetation plan for the onshore work area (Chevron 1986) prepared by Chevron after the EIS/EIR as part of their final development plan (FDP) as required by Santa Barbara County permit conditions. This plan was developed in consultation with Santa Barbara County and refinements were made in response to comments from the Santa Barbara County Resource Management Department, Energy Division. Further developments and refinements of the approved plans were made in the field during their implementation. This was generally done with the input of engineers, environmental staff, and consultants employed by Chevron, and input from environmental staff, monitors, and consultants representing Santa Barbara County. Environmental monitoring to ensure compliance with permit conditions was done according to a detailed environmental quality assurance program (EQAP) employing an onsite environmental coordinator (OEC), general environmental monitors, and resource specialists such as biologists.

Conveying the stream water across the construction area in a flume minimized the construction-related sediment that could have fallen into the water. At the downstream side of the ROW, silt traps were installed prior to construction. These traps typically consisted of a combination of two staggered rows of straw bales anchored with rebar and a curtain of commercial silt fence fabric keyed into the stream bottom at the downstream side of the straw bales. These sediment traps served as the primary means of sediment control after the trench had been backfilled and the flume removed. Accumulated sediment was removed from the upstream side of the traps by hand and disposed of away from the creek. It was found that the silt fence fabric by itself was ineffective. The fabric rapidly became clogged with sediment, causing stream flow to back up until the water began

228

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

to flow around, over, or under the fence. The silt fences required regular monitoring and maintenance. The staggered straw bales were relatively effective and simple to maintain, provided that the flow was not too great. At higher flows, the sediment traps acted like a dam, retarding the flow until the water backed up and cut around the sides of the bales, eroding the banks. Sediment would drop out of the water behind the dam. Eventually this problem was solved by installing one or more L-shaped overflow pipes that would take the cleaner surface water from behind the dam and convey it under the hay bales and silt fence. The sediment traps were left in place until after the first rainstorms following construction. Prior to their removal, the accumulated sediment was removed and disposed of away from the creek bed. Accumulation of water in the trench created a problem at several crossings. When pumped from the trench, the discharged water tended to be sediment-laden, and the concentrated pump discharge caused local scouring of the streambed. In cases of very low discharge rates, the sediment was controlled by silt traps such as those described above. At certain streams, groundwater accumulation necessitated pump discharge rates that exceeded the capacity of the sediment control system. In these cases, one or two swimming-pool-sized "baker" tanks were used to receive discharge and to act as sediment traps by allowing sediment to settle out of the relatively calm water. Sediment-free water was then siphoned from the top of these tanks and returned to the stream bed. An alternative way of handling the higher rates of accumulated groundwater was to discharge it through sprinklers to upland habitats draining into the creek. At one crossing, buried marsh sediments yielding abundant groundwater were encountered during trenching. This water was laden with hydrogen sulfide and fine sediment. Observation at other stream crossings indicated that the material was toxic to invertebrates. To maintain workable conditions in the trench, several shallow dewatering wells were installed and cased with perforated PVC plastic drain pipes. The discharge from these wells was too copious for the above-described systems to handle. The short distance of stream between the construction site and the shore was considered extremely sensitive to impact because it contained three aquatic species that are candidates for federal listing (tidewater goby, red-legged frog, and southwestern pond turtle) and had very little flow at the time of construction. To protect the stream and its biota, the water was piped to the nearby beach and discharged onto the sand, where filtering by the sand and dilution in the ocean water would minimize adverse ecological effects.

Stream-Bank Stabilization

After completion of welding and testing, the trench was backfilled, and the crossing was recontoured to its approximate preconstruction configuration. Original specifications called for armoring stream banks below the normal high water level with a three dimensional nylon "geotextile fabric." This armoring was intended to protect the banks and cause sediment to accumulate in the mesh, where it would ultimately be stabilized by vegetation. In practice, the fabric proved to be relatively difficult to install, especially where it had to be keyed in to the stream bottom. There were concerns about its longevity, the aesthetic impact of exposed portions of the mesh, and the possibility of its being washed out by floodwater and causing downstream problems (e.g., clogging culverts). There was also a concern that the mesh would cause girdling of roots of willows and other woody plants expected to colonize the banks. As a result, this material was removed from several drainages. Where it was left in place, the material functioned adequately in low-flow situations where the drainage was broad with little gradient. However, these situations are the least prone to erosion. Elsewhere, it was heavily damaged by seasonal runoff or by livestock. Based upon this experience, we would not recommend the use of this material in a similar environment, especially in areas frequented by cattle. Water bars (earthen berms graded across the ROW at a slight angle to the topographic contour) were installed at intervals along the steeper slopes to divert runoff from the ROW into adjacent vegetated areas. The spacing between water bars depended upon the steepness of the slope. On the steepest slopes, they were installed at 25-foot intervals. These bars effectively prevented the concentration of runoff and consequent erosion along the ROW but did require periodic repairs after winter storms. Damage caused by livestock also required repair. Gullying occurred at the edge of the ROW if the berms were not continued a slight distance beyond the ROW into the adjacent vegetation. Jute netting was installed on steep slopes above the high water mark. This fabric was intended to retard washing of the exposed soil and backfill from the ROW into the creek. It was also expected to facilitate the establishment of vegetation by trapping and holding seeds in place, as well as by holding the soil. Proper installation requires that the fabric be placed in intimate contact with the soil surface, otherwise it can be undermined by runoff and become ineffective. Damage caused by cattle was extensive. Where properly installed and undamaged by cattle, the jute appeared to be an effective erosion-control measure. Postconstruction revegetation on the jute-netted slopes has been relatively slow,

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

229

however, leading to speculation that some property of the jute might have been inhibitory to seed germination or establishment. We believe that this could as easily have been a function of the poor soils and steep slopes where the jute was used. We would recommend establishment of appropriate control areas in subsequent applications to resolve this issue and to better evaluate the effectiveness of the material.

Minimize Vegetation Removal

Santa Barbara County permit conditions required construction activities to be confined to a 100-foot wide ROW. Narrowing the ROW in sensitive habitats and avoiding sensitive resources such as trees wherever feasible was also required. Preconstruction walk-throughs were conducted by Chevron's consultant biologists and a Chevron inspector familiar with pipeline construction methods. Trees and large shrubs that would unavoidably be removed by construction were enumerated by species for the purpose of developing site-specific replanting specifications. In the process, possibilities for minimizing impacts were discussed, and sensitive vegetation was identified, prioritized, and flagged for avoidance. Later, trees were protected by surrounding them with heavy lath and wire snow fencing. Because we attempted to err on the side of saving trees, we identified and flagged the maximum number of plants that could possibly be avoided by construction, using optimistic assumptions. It was later determined that removal of some marked trees was unavoidable, and these were unfenced after notification of the on-site environmental coordinator. The snow fence proved to be very effective both in designating trees or resources to be avoided and in preventing the plants from being buried by spoil or damaged by heavy equipment operating in their vicinities.

potted using native soil. The plants tended to fragment into rooted divisions during the excavation and were thus placed three divisions to a pot into 15-gallon plastic pots, each of which was color-coded according to source locality. Initially, the plants were maintained in the field near the ROW where they were protected by hog wire and shade cloth. Subsequently, they were moved to a local nursery where they were maintained outdoors. Normally a summer-dormant, drought-deciduous species, the plants had been nearly dormant when salvaged. However, many plants became active again and flowered in response to periodic watering during the summer months. There was a relatively high but unquantified survival of salvaged plants in the nursery. Replanting was accomplished during spring 1987 at their original localities in chicken-wire exclosures. Thirty-eight percent of the outplanted plants, including several plants from each local population, were surviving when last monitored (July 1988). These appear to be fully established. Also undertaken were transplants of giant stream orchids (Epipaclis gigantea) and scarlet monkey flower (Mimulus cardinalis). Both of these plants are widespread species but were singled out for this treatment because of their uncommonness in the project area. Both species occurred under the riparian canopy in relatively moist sandy soil adjacent to the bed of Canada del Cojo, one of the larger streams crossed by the ROW. The plants were dug up and immediately replanted in one of two preselected open sites outside of the ROW, one upstream and one downstream. Finding sites with the proper attributes that were not completely covered with vegetation was perhaps the most difficult aspect of this task. Excavating the root system of the monkeyflower proved to be impractical and the effort was abandoned after only a few were transplanted. To my knowledge, none of these survived. About a dozen of the stream orchid were moved, and some of these were surviving as of spring 1988.

Preserve Rare Plants

Protect Sensitive Wildlife and Aquatic Species

During preproject walk-throughs, seven local populations of Hoffmann's nightshade (Solanum xanli var. Hoffmannii), a California Native Plant Society List 4 species endemic to coastal Santa Barbara County, were located. Hoffmann's nightshade is a sprawling subshrub that clambers through other plants and roots at the nodes. Its characteristics led us to believe that it could be successfully transplanted (Mulroy and others 1984), and transplanting was incorporated into a permit condition by Santa Barbara County. The local populations were therefore flagged in the field for later salvage. In early summer 1986, prior to construction, Hoffmann's nightshade plants were cut back, dug up, and

The efforts to maintain stream flow and to protect instream water quality, described above, were directed at preserving local populations of the above-listed sensitive aquatic species. A permit requirement that construction be restricted to the low-flow season was directed at making these protection measures feasible. A related condition was placed on the project to delay construction until after the breeding period of a number of regionally rare and declining songbird species known or expected to breed in some of the habitats. The net effect of these two conditions was to leave a very narrow "window" for construction at stream crossings. Chevron approached this constraint by prioritizing stream crossing construction, having special crews work on the stream crossings.

230

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

These crossings were later tied in as the main construction spread traveled from west to east. Survival of the tidewater goby populations was of special concern since, during most of the year, the fish are located in tiny lagoons near the stream mouths, where they would be vulnerable to local extirpation. Recolonization could only occur from other populations via the ocean and there had been no documented evidence that this could occur. For these reasons, sampling to monitor potentially affected goby populations was conducted before, during, and after construction as part of the Environmental Quality Assurance Program (EQAP). This sampling is described in a poster session at these meetings authored by Thompson, Dunlap, and Dungan.

upon the cherry fruits extensively. The work of pack rats or wood rats (Neotoma) in collecting both scats and seed was taken advantage of in respect to this seed-collection attempt. For most of its length, the ROW passed through active cattle ranches. Protection from grazing livestock was an important factor in the design of the revegetation program and was important in determining the success of the tree and shrub plantings. Box fencing was most effective in excluding cattle from the ROW at stream crossings, but they prevent up and downstream movement of cattle in the streambed, and for this reason landowners strenuously objected to their installation at most locations. Where properly installed, wing fences (fences constructed across the ROW and angling downslope for varying distances off the ROW) inhibited cattle from using the ROW for access to stream crossings and, in such areas, tree and shrub plantings fared substantially better than in areas in which fencing was absent or ineffective. In the latter cases, cattle tended to congregate in the ROW at the stream crossings and restoration plantings, and erosion control measures were heavily affected, requiring considerable maintenance and repair. In upland habitats outside of the wing-fenced areas, a two-tiered exclusion system designed by one of the authors was used to exclude cattle, deer, and small mammals from shrub and Hoffman's nightshade plantings. The plantings were clustered within an area enclosed with standard three-strand barbed-wire fencing designed to exclude cattle. The individual plantings were protected from deer and small mammals by individual cages fabricated from chicken wire. These cages were about 30 inches in diameter and about 54 inches from top to bottom. The bottom 12 to 15 inches of the cage was buried below ground to provide protection from pocket gophers (Thomomys bottae). The timing of revegetation activities is critical. Plans had been to complete pipeline construction and restoration activities prior to the onset of the November-toApril rainy season, in order to give the plantings and reseeded areas the maximum amount of time to become established before the summer drought. The project, unfortunately, fell behind schedule, and revegetation activities were not initiated until January 1987. Tree and shrub plantings were not accomplished until later that spring, when runoff had diminished to low levels and soils in the riparian areas away from the immediate channel had dried out. As a result, establishment of riparian trees and shrubs from cuttings and transplants was poor, except in areas within a few feet of the existing channel. Considerable reseeding of adjacent shrub-dominated upland habitats was required as well, necessitating additional custom collections of native seed from the immediate vicinity of the ROW.

Restore Native Vegetation

A detailed description of measures to revegetate the stream sides and adjacent slopes is beyond the scope of this paper. However, I would like to briefly mention a few things germane to the restoration of streambeds and riparian zones in particular. First, willow cuttings taken from trees immediately adjacent to the ROW were a rapid and inexpensive means of hastening the restoration of a tree cover over the stream. Generally, the use of dormant wood is preferred; however, cuttings taken in late spring survived. There was a high rate of survival of cuttings where soil moisture was adequate (i.e., near the streambed) even though they were not taken and planted until late spring. Second, to hasten the stabilization of the streambed and recovery of marsh plants such as cattails in severely disturbed portions of the ROW, a simple method for salvage and replanting was developed for this project. This method involved digging individual clumps of cattails by hand out of the ROW prior to clearing, and storing them in plastic wading pools for later replanting. The wading pools can be located in shaded parts of the streambed adjacent to the ROW and watered periodically by construction monitors or inspectors. The ultimate success of this method remains untested because an errant bulldozer destroyed the plants and tub. Reestablishment of cattails from seed and rhizomes in the salvaged topsoil was rapid, however. An example of a creative approach to seed collection was taken by one of the authors after a seed-collection contractor missed the brief period of seed availability of holly-leaf cherry or islay (Prunus ilicifolia), a large shrub or small tree forming coppices on steep slopes leading into several streams. A considerable amount of locallycollected seed was required for this species, which was one of the most important large shrubs along the ROW. One of the authors was able to collect nearly the entire required amount from the scats of coyotes which had fed

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

231

Conclusions and Recommendations


The maintenance of stream flow during periods of low flow proved easier than did the removal and redistribution of groundwater which accumulated in the pipeline trench. A trial-and-error approach was used to determine the best method for each drainage where this problem arose. For similar projects involving excavation through riparian zones, we would recommend that an estimation of groundwater quantity and quality and tactics for dealing with this excess water be determined prior to construction. As was the case with the Point Arguello pipeline project, appropriate site-specific solutions will have to be developed for this problem. In streams larger than those encountered in this project, reinjection of the groundwater may be feasible. Use of geotextile fabric for stream bank stabilization did not appear appropriate and was discontinued. We believe that the revegetation of stream banks provides a superior degree of bank stabilization, without the potential drawbacks of the synthetic material. Erosion was noted where water bars led into steep and poorly vegetated areas at the edge of the ROW. Otherwise, water bars were effective but required periodic repair of damage caused by storms or livestock. Investigations of the effects of jute netting on plant growth are recommended. However, our experience suggests a possible inhibitory effect on revegetation. Snow fencing was highly successful in protecting trees from burial and incidental damage by equipment operators. A number of suggestions are made above for enhancing the reestablishment of riparian vegetation. An important step in much of the West is to restrict the access of livestock to recovering stream corridors. These areas are most vulnerable to damage from livestock during summer, when the animals tend to congregate in riparian areas and may do irreparable damage to young plants. Finally, it is important that techniques to protect riparian zones from construction impacts be scrutinized and refined. Every attempt should be made to distinguish causal relationships from coincidences when the success or failure of mitigations is assessed. Ideally, this will involve rigorous experimental comparisons between treated areas and control areas.

Acknowledgements
To make a list will ensure that some are left out, nevertheless the authors would like to acknowledge the efforts of the following: We thank Roz Muller, John Tiffany, Dennis Kennedy, Ted Potter, George Boscoe, Peter Catton, and Anthony Demes of Chevron U.S.A. Inc. or Chevron Pipe Line Company; George Welsh, of Ecology and Environment, Inc., and several persons from Santa Barbara County Resource Management Department, Energy Division, including Peter Cantle, Mary Ann Scott, and Tom Lagerquist for their creativity, patience, willingness to listen, cooperation, and hard work.

References
Arthur D. Little (ADL), Inc. 1984. Point Arguello field and Gaviota processing facility area study and Chevron/Texaco development plans EIR/EIS. Prepared for County of Santa Barbara, U.S. Minerals Management Service, California State Lands Commission, California Coastal Commission, California Secretary of Environmental Affairs. 585p. Available from Santa Barbara County, Resource Management Department, Energy Division. Chevron Pipe Line Company. 1986. Restoration, erosion control and revegetation plan, onshore work area. Prepared by Chevron Pipe Line COmpany, Concord, CA for Point Arguello Pipeline Company and Point Arguello Natural Gas Line Company for submission to Santa Barbara County, Resource Management Department, Energy Division. Mulroy, Thomas W., Thompson, R.; Hochberg, M.; Collins, P.; and Lehman, P. 1984. Terrestrial and freshwater biology. With contributions from W. Ferren and S. Junak. Technical Appendix J of ADL, Inc. 1984. 224p. Santa Barbara County. 1982. Local Costal Plan. Santa Barbara County Resource Management Department, Comprehensive Planning Division. Santa Barbara, CA. Santa Barbara County. 1984. Final permit actions: Chevron Pt. Arguello/Gaviota oil and gas development project. Approved by Santa Barbara County Board of Supervisors, Santa Barbara, California. December 21, 1984. Santa Barbara County, Resource Management Department, Energy Division.

232

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

SESSION F: WILDLIFE I-MANAGING FOR SELECTED SPECIES


One of the more obvious signs that riparian habitat has all but disappeared in California is the number of endangered species whose existence depends on it. Because of the legal protection accorded to endangered species, they have become a major reason that riparian habitat is being protected and restored. Riparian habitats that harbor endangered or threatened species or even "species of special concern," suddenly become impediments to development of marinas, rip-rapping projects, and rows of streamside condominiums. Then finding out as much as possible about the requirements of these species becomes imperative, so that the minimum amount of habitat can be set aside for their protection. Under these conditions, money usually becomes available for the study of some of the more obscure but fascinating species that inhabit riparian areas. The twelve papers presented here deal with seven discrete animal groupsstoneflies, a beetle, salmonid fishes and four species of riparian birds. The beetle and all of the birds are considered to be at some level scale of concern over possible extinction. Protecting such fascinating species is reason enough to save riparian habitats. Peter B. Moyle University of California, Davis

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

233

STONEFLY (PLECOPTERA) FEEDING MODES: VARIATION ALONG A CALIFORNIA RIVER CONTINUUM 1


Richard L. Bottorff and Allen W. Knight2

Abstract: The distribution of Plecoptera along a California river was used to test several predictions of the River Continuum Concept about how functional feeding groups should change along a stream's length. Stoneflies were collected from stream orders 1-6 (123 km) of the Cosumnes River continuum in the central Sierra Nevada. The 69 stonefly species collected were separated into three functional feeding groups 26 shredders (detritivores), 39 predators (engulfers), and 4 scrapers. Stonefly shredders and predators reached maximum species richness at middle stream orders (3-5). The relative abundance of shredders was highest in the headwaters, then declined with increasing stream order. Predator relative abundance exhibited an inverse trend to shredders. Few scrapers were present, but these were most abundant at middle stream orders; they were almost entirely absent from the headwaters. These trends in diversity and relative abundance of functional feeding groups agree with predictions of the River Continuum Concept and emphasize the importance of riparian vegetation in the structure and function of Sierra Nevada streams. The type of terrestrial plant community through which the stream was flowing did not appear to affect the diversity or relative abundance of shredders.

more sunlight can reach the substrate and enhance instream primary production. Scraper organisms, which graze algae and fine particles from the substrate, are predicted to be relatively abundant in the middle reaches of streams (orders 4-6) (Vannote and others 1980). Maximum biotic diversity is predicted to occur in middle stream orders where environmental variation is greatest. The RCC has been an important catalyst for recent stream research throughout the world (Minshall and others 1985, Statzner and Higler 1985), and the validity of its tenets and predictions are currently being tested. However, the predictions of the RCC and the importance of riparian vegetation to the macroinvertebrate community have not been studied in Sierra Nevada streams, despite the potential value for watershed and forestry management practices. Stoneflies are ideal organisms for testing some of the RCC predictions about how FFG's change along a river continuum. First, stoneflies are common components of the aquatic benthos of most streams worldwide, often with many species and individuals present at any particular location. Since they extend in distribution from headwaters to far downstream, long distance changes in faunal composition, FFG's, and diversity can be studied. By concentrating on a single aquatic insect order, a relatively complete taxonomic list can be obtained at the species level, eliminating the uncertainties of incomplete identifications. In keeping with the idealized, undisturbed stream of the RCC, stonefly nymphs are almost exclusively restricted to unpolluted running waters. In contrast with some aquatic insect groups, stonefly nymphs belong predominantly to only two of the six main FFG's (Merritt and Cummins 1984) shredders (detritivores) and predators (engulfers). This basic feeding dichotomy is reflected in the morphological differences of their mouthparts, and also in the two taxonomic groups of the suborder Arctoperlaria (Zwick 1973), the Euholognatha (mainly shredders) and the Systellognatha (mainly predators). A few stoneflies are classed as scrapers (the Brachypteriginae), but very few are classed as collectors (Merritt and Cummins 1984). Stoneflies are also good study organisms because most of the food resources needed during their life cycle are obtained as nymphs in the stream habitat, many adults being short-lived and not feeding (the Systellognatha). Because of the advantages of stoneflies, we used

The River Continuum Concept (RCC) was proposed to explain the structural and functional changes which occur in stream communities, especially to streams flowing naturally through undisturbed forested watersheds (Vannote and others 1980). Inputs of organic detritus to the stream and detrital processing by functional feeding groups (FFG's) of aquatic invertebrates are central themes of this concept. The RCC predicts that shredder organisms, which consume large particles of organic detritus (>1 mm) that has accumulated on the stream substrate, should be relatively abundant near the headwaters of a stream, then decrease downstream as the stream widens and the relative importance of coarse detrital inputs from terrestrial vegetation and upstream habitats decline. The rationale is that the smaller streams in the upper reaches of a river continuum (stream orders 1-3) are greatly influenced by riparian vegetation, the overhanging plant canopy supplying large amounts of organic detritus, but shading and limiting instream primary producers. As the stream widens downstream and detrital inputs from streamside plants become less important,

1Presented 2

at the California Riparian Systems Conference; September 22-24, 1988; Davis, California.

Graduate Student and Professor, Department of Land, Air, and Water Resources, University of California, Davis, California 95616

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

235

them to test several predictions of the RCC in an undisturbed California stream. The main objectives of the study reported here were to determine (1) if the relative abundance of stonefly shredders declines with increasing stream order (1-6), (2) if the relative abundance of stonefly scrapers increases with stream order to a maximum in middle stream orders (4-6), and (3) if the maximum diversity of FFG's occurs in the middle reaches. The RCC predicts little change in predator relative abundance along the continuum; however, because stoneflies are largely either shredders or predators, these two FFG's will demonstrate inverse trends. A final objective of the study was to observe if the diversity and relative abundance of the stonefly FFG's were influenced by the type of terrestrial plant community being traversed by the stream.

In contrast to most large rivers on the western slope of the Sierra Nevada, the Cosumnes River has no mainstream reservoirs; therefore, water temperature and discharge change naturally with the seasons. Because of the largely undisturbed watershed and lack of mainstream reservoirs, the Cosumnes River possesses the pristine conditions that formerly existed in many of the larger streams of the western Sierra Nevada and should be close to the typical stream defined in the RCC.

Methods
The distribution of stoneflies was determined for a 123 km reach of the Cosumnes River, extending from the headwaters (2249 m elevation) to the lower reaches (27 m elevation). Stonefly nymphs and adults were collected from nine sites in this study segment, which included stream orders 1-6 and four main terrestrial plant communities. Stoneflies were abundant at all nine sites of this study, including sites near sea level, but they were absent from the extreme lower reaches of the continuum ( > 123 km, stream orders 7) in the Sacramento-San Joaquin Delta (California Department of Water Resources 1981). Plecoptera nymphs were collected monthly (19801982) from both riffles and pools at each of the nine sites using a kick screen (0.9 mm mesh size). Most identifications were made to genus or species level; a few nymphs could be identified only to family level (some Capniidae, Leuctridae, and Chloroperlidae). The inability to identify all nymphs to species was not thought to greatly affect this study because identification to genus or family was often sufficient to determine the proper FFG. During this study, a total of 26,000 Plecoptera nymphs were collected and identified from the nine sites. Stoneflies were placed into FFG's based primarily on Merritt and Cummins (1984), plus the examination of gut contents for some stonefly nymphs when their feeding status was uncertain. Plecoptera adults were collected in all seasons (19801986) using three methods: general search, timed sweepnet, and slit traps. General search along the streamside included examination of those substrates likely to contain adults, but difficult to adequately sample with a sweep net. In winter at high elevations, stoneflies were hand collected from snowbanks along the stream. Stream segments buried under snow were dug open, then adults collected as they emerged. Timed sweepnet collections (10 - 120 min) were made at all sites (biweekly 1981-1983, irregularly thereafter) to collect adults on streamside vegetation. Adults were also collected in slit traps (1981-1982) designed primarily to capture newly emerged Plecoptera (Kuusela and Pulkkinen 1978). These traps were placed on the stream

Study Stream
Plecoptera feeding modes were studied along one river continuum, the Cosumnes River and its North Fork, located in the central Sierra Nevada east of the city of Sacramento, California. The study stream flows westward 163 km from its headwater springs at 2249 m elevation to nearly sea level in the plains-like Central Valley, where it joins the Mokelumne River and enters the Sacramento-San Joaquin Delta estuary. Large physical gradients exist between the headwaters and lower reaches of the Cosumnes River, especially in mean discharge (0.007 - 13.8 m3/sec), channel slope ( 12.9 0.06 percent), substrate (-8 to +3 phi scale), summer water temperature (4 - 28C), and stream order (1-6). It is an unpolluted softwater stream (specific conductivity = 20-115 Mmho/cm), with dissolved inorganic solids increasing only slightly between the headwaters and lower reaches. The Cosumnes River drains a largely undisturbed, forested watershed and traverses four major terrestrial plant communities as it descends the large elevation gradient of the Sierra Nevada (Griffin and Critchfield 1976) red fir forest (> 2000 m elevation), mixed conifer forest (730-2000 m elevation), foothill woodland and chaparral (40-730 m elevation), and valley grassland (< 40 m elevation). Forests above 730 m elevation are primarily coniferous (Abies, Libocedrus, Pinus, and Pseudotsuga); forests below 730 m elevation are primarily broad leaf forms (Aesculus, Arctostaphylos, Ceanothus, and Quercus). In addition to these four terrestrial plant zones, the stream is bordered along its entire length by a narrow band of true riparian plants (Alnus 2 species, Populus 2 species, Salix many species, and other trees, shrubs, and herbs). The entire Cosumnes River drainage basin lies below tree line.

236

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

bank just above water level immediately adjacent to the stream and operated continuously, day and night. Using these three methods, a total of 12,000 adults were collected from the nine sites, identified to species, and separated into their FFG's. In addition, stonefly exuviae were collected from rocks and vegetation along the stream banks and were identified to species for three families (Perlidae, Perlodidae, and Pteronarcyidae). Since any collection of adult stoneflies along a river continuum represents a mixture of species derived directly from the river, plus those dispersing in from nearby smaller tributaries, an effort was made in this study to distinguish between these two faunas. To separate these two faunas, we (1) collected both nymphs and adults along the continuum, (2) associated many nymphs with adults, and (3) collected less intensively the nymphs and adults from side tributaries. The separation of adventitious species from the true river continuum fauna was important because our study was limited to FFG changes along a single continuum. Stream order was used to define the position along the continuum. Stream order locates a stream segment within the drainage network (Strahler 1957) and is determined by the number and size of upstream tributaries. Headwater streams lacking inflowing tributaries are first order; higher order streams result from the confluence of two lower order streams (second order from the confluence of two first order streams; third order from the confluence of two second order streams, etc.) In this study of the Cosumnes River continuum, stream order was determined using U. S. Geological Survey topographic maps (1:24,000 scale), and all intermittent streams were included within the drainage network.

shredders and slightly more predators than the other two areas. These slight differences were sampling artifacts caused by comparing the stonefly faunas of two large areas with a linear sample of one river continuum. Because we restricted our study to a single river continuum, stoneflies occurring in low order, low elevation streams in the Cosumnes River basin were excluded from the study, and these low order streams should contain higher proportions of shredders. Species richness of stonefly shredders and predators was highest in middle stream orders (3-5) of the Cosumnes River continuum, with shredders reaching a maximum of 20 species at stream order 3 and predators reaching a maximum of 23 species at stream order 4 (fig. 2A). Near the headwaters and in the lower reaches of the stream, fewer species of both FFG's were present. At stream order 7, species richness was zero. These diversity trends in shredders and predators agree with the RCC prediction of maximum biotic diversity at middle stream orders. Stonefly scrapers were much less abundant than shredders and predators, but increased from none to three species along the continuum.

Results
A diverse stonefly fauna occurred in the Cosumnes River continuum. After excluding adventitious species, a total of 69 species, 36 genera, and all 9 North American families of stoneflies were collected in this study. Of these 69 species, 26 were shredders, 39 were predators, and 4 were scrapers (Brachypterinae: Taenionema). Four Chloroperlidae genera (Alloperla, Bisancora, Kathroperla, and Paraperla) may fit at least partially into the collector FFG; however, because few species were involved and little was known of their feeding mechanisms, the collector FFG was not considered further in this study. The proportions of stonefly shredders, predators, and scrapers collected from the Cosumnes River continuum (fig. 1) were generally similar to the stonefly faunas of North America and California (Stark and others 1986), although the Cosumnes River fauna had slightly fewer Figure 1 Proportion (percent) of three functional feeding groups composing the stonefly faunas of North America, California, and the Cosumnes River continuum. 237

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 2 Variation in species richness and relative abundance of stonefly functional feeding groups (FFG's) in stream orders 1-6 of the Cosumnes River continuum, California: (A) species richness of FFG's, (B) relative abundance (percent) of FFG's determined from the total stonefly species collected (adult + nymph data), (C) relative abundance (percent) of FFG's determined from the total number of stonefly adults collected, (D) relative abundance (percent) of FFG's determined from the total number of stonefly nymphs collected.

The relative abundance of stonefly FFG's along the Cosumnes River continuum was examined in three ways: (1) using the number of stonefly species collected (fig. 2B), (2) using the number of adult specimens collected (fig. 2C), and (3) using the number of nymph specimens collected (fig. 2D). All three data sets gave similar results (fig. 2B- D); shredder relative abundance decreased as stream order increased from 1 to 6, while predator and scraper relative abundance increased. Although the decrease in shredder relative abundance with increasing stream order occurred at slightly different rates in each data set, shredders were always most abundant (> 60 percent) near the headwaters and much less abundant in the lower reaches. Because of the basic FFG dichotomy in stoneflies, predator relative abundance ex-

hibited an inverse trend to that of shredders. Scraper relative abundance was low (< 15 percent) at all stream orders and in all data sets, but did increase slightly in middle orders (3-6). Scrapers were especially sparse or absent from the Cosumnes River headwater region (orders 1-2). The species richness and relative abundance of stonefly FFG's did not appear to be greatly affected by the type of terrestrial plant community through which the stream flowed (fig. 3A-B). Changes in species richness and relative abundance occurred gradually between sites, rather than exhibiting discontinuities at vegetation zone boundaries. The maximum number of shredder and predator species occurred within the mixed conifer forest zone (fig. 3A).
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

238

and at high elevations where riparian vegetation is sparse (Busch and Fisher 1981). The concept was originally formulated for, and appears most applicable to, streams flowing through undisturbed, forested watersheds. Since the Cosumnes River has an undisturbed, forested watershed, the predictions of the RCC would be expected to apply. The results of this study agree with the RCC predictions that (1) shredders should be most abundant near the headwaters and decrease in relative abundance downstream, and (2) scrapers should be most abundant at middle stream orders. These changes in stonefly FFG's along the Cosumnes River continuum, plus observations of the watershed and stream canopy vegetation, strongly suggest that streamside vegetation (main terrestrial plant communities + true riparian plants) had an important organizing role on the stream biota, both as a source of organic detritus and as a controller of sunlight entry into the stream. It is likely that other streams traversing the heavily forested watersheds on the western slope of the Sierra Nevada are similarly influenced by riparian vegetation and demonstrate similar FFG changes with the Cosumnes River. Possible exceptions are those Sierra Nevada rivers extending above tree line where inputs of organic detritus may be sparse. Additional tests of the other RCC predictions should be made on Sierra Nevada streams using the entire macroinvertebrate community or various subsets. Few other studies have examined stonefly feeding groups along a stream continuum. In contrast to our results, Knight and Gaufin (1966) found the proportion of predatory stonefly species increased with elevation in a Colorado drainage, while the proportion of vegetarian stonefly species decreased. These trends may be caused if the high elevation (> 10,000 feet) Colorado stream was above tree line and lacked sufficient inputs of organic detritus from riparian vegetation. Brink (1949) also found carnivorous stoneflies reached high dominance values on a mountain ridge in northern Sweden; however, these waters contained few plants and had little organic detritus. He found that phytophagous stonefly species were dominant in springs, trickles, and small eutrophic forest streams, all of which contained large amounts of organic detritus. These studies suggest that stonefly feeding modes vary directly with the food resources available in the stream habitat. Stonefly scrapers were especially sparse or absent from the headwaters of the Cosumnes River continuum. We believe this is caused by the combination of heavy shading by riparian vegetation and by deep snow accumulations which bury the stream for many months in winter- spring. However, this possibility should be tested with other aquatic insect groups which naturally include more species of scrapers. Although the RCC predicts maximum biotic diversity at middle stream orders, it was somewhat surprising to 239

Figure 3 Variation in species richness and relative abundance of stonefly functional feeding groups (FFG's) by elevation and terrestrial plant community along the Cosumnes River continuum, California: (A) species richness of FFG's, (B) relative abundance (percent) of FFG's determined from the total stonefly species collected (adult + nymph data). Plant community names: VG = valley grassland; FW = foothill woodland and chaparral; MC = mixed conifer forest; and RF = red fir forest. Stream orders are given to show their relation to elevation and the terrestrial plant communities.

Discussion
Since the RCC was first proposed (Vannote and others 1980), its worldwide applicability to all streams has been debated (Winterbourn and others 1981; Barmuta and Lake 1982; Minshall and others 1983, 1985; Statzner and Higler 1985), and certain modifications and exceptions have been noted, especially for streams in deserts
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

find this was also true for the different stonefly FFG's in the Cosumnes River continuum (fig. 2A). Originally, we expected shredder diversity to be highest near the headwaters where their food source of coarse organic detritus should have been most abundant. However, shredder diversity was maximum at stream order 3 and remained high at stream orders 4-5, suggesting that some factor other than detrital quantity controls diversity. Possibly, a favorable diversity of substrate or detritus promotes shredder diversity in middle stream orders. Maximum shredder diversity did occur slightly closer to the headwaters (order 3) than did maximum predator diversity (order 4-5). Many studies have shown the importance of streamside vegetation to stream ecosystems, and it is well known that the type of terrestrial vegetation greatly affects the amount and timing of detrital input to streams, the decay rate of organic detritus, and the food preferences of shredders (see references in Hynes 1975, Minshall and others 1985). Some research suggests that the type of terrestrial vegetation through which the stream flows influences the composition of the benthic stream community (Ross 1963, Donald and Anderson 1977, Wiggins and Mackay 1978, Culp and Davies 1982, Molles 1982). Although the Cosumnes River continuum passed though four main terrestrial plant communities, no major discontinuities in species richness or relative abundance of stonefly shredders or predators were evident near vegetation zone boundaries. Species richness was maximum within the mixed conifer forest zone for both FFG's, but the gradual change in diversity across zone boundaries suggests vegetation type had little influence. Also, the decrease in shredder relative abundance with increasing stream order occurred uniformly across vegetation boundaries, suggesting that shredders were responding to the decreasing quantity of coarse organic detritus, rather than to the type of terrestrial vegetation. The lack of FFG discontinuities at terrestrial vegetation zone boundaries may be explained by (1) the relative contribution of organic detritus coming from the main terrestrial plant communities versus that from the narrow border of true riparian vegetation, and/or (2) the downstream transport and mixing of organic detritus by the stream. Sierra Nevada streams are typically bordered by a few true riparian plant species which are adapted to the moist soil conditions near the stream. These riparian species often parallel the stream over much of its length, and extend their distribution through several terrestrial plant zones. Leaves shed from these true riparian species are known to quickly become preferred food items for shredders (Hynes 1975). If a major portion of the organic detritus originates from the true riparian species, then discontinuities in shredder relative abundance and diversity would not be expected to occur at the main terrestrial vegetation boundaries. The

relative importance of the main terrestrial plant communities and the true riparian vegetation as detrital sources should be investigated further in Sierra Nevada streams. Although stoneflies fit into two main feeding groups, shredders and predators (Merritt and Cummins 1984), much more research on food habits is needed for many genera and species. To accurately place a stonefly within one or more FFG's, it is not sufficient to just examine gut contents; in addition, observations of mouthpart morphology and feeding behavior are needed. Determination of FFG is complicated by food habit variations with stonefly age, size, and environment (Hynes 1976). Despite these uncertainties, few stoneflies seem to have evolved scraper and collector feeding modes, possibly as Wiggins and Mackay (1978) have speculated, because other aquatic insect orders (Ephemeroptera, Trichoptera, and Diptera) have completely exploited these feeding methods. Filter feeding is noticeably absent in Plecoptera nymphs, but is common in these other aquatic insect orders. An interesting outcome of this study was that data from a small subset of the entire stream macroinvertebrate community could be used to test and support several predictions of the RCC, which was proposed to explain longitudinal changes in stream ecosystems and the entire macroinvertebrate community. Wiggins and Mackay (1978) used Trichoptera to test RCC predictions of FFG's and to compare differences between streams in eastern and western North America. The use of other taxonomic subsets to test the RCC could offer additional important insights into stream processes.

Acknowledgments
This study was supported by a Research Assistantship and by Jastro-Shields and Graduate Research Grants, from the University of California, Davis.

References
Barmuta, L.A.; Lake, P.S. 1982. On the value of the River Continuum Concept. New Zealand Journal of Marine and Freshwater Research 16:227-229. Brinck, Per. 1949. Studies on Swedish stoneflies (Plecoptera). Opuscula Entomologica Supplementum 11:1-250. Busch, David E.; Fisher, Stuart G. 1981. Metabolism of a desert stream. Freshwater Biology 11:301-307. California Department of Water Resources. 1981. SacramentoSan Joaquin Delta water quality surveillance program, 1979. Vol. III. California Department of Water Resources, Sacramento, CA. 118 p.

240

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Culp, Joseph M.; Davies, Ronald W. 1982. Analysis of longitudinal zonation and the River Continuum Concept in the Oldman - South Saskatchewan River system. Canadian Journal of Fisheries and Aquatic Sciences 39:1258-1266. Donald, David B.; Anderson, R. Stewart. 1977. Distribution of the stoneflies (Plecoptera) of the Waterton River Drainage, Alberta, Canada. Syesis 10:111-120. Griffin, James R.; Critchfield, William B. 1976. The distribution of forest trees in California (with supplement). Res. Paper PSW-82. Berkeley, CA: Pacific Southwest Forest and Range Experimental Station, Forest Service, U.S. Department of Agriculture; 118 p. Hynes, H. B. N. 1975. The stream and its valley. Verh. Internat. Verein. Limnol. 19:1-15. Hynes, H. B. N. 1976. Biology of Plecoptera. Annual Review of Entomology 21:135-153. Knight, Allen W.; Gaufin, Arden R. 1966. Altitudinal distribution of stoneflies (Plecoptera) in a Rocky Mountain drainage system. Journal of the Kansas Entomological Society 39:668-675. Kuusela, Kalevi; Pulkkinen, Hannele. 1978. A simple trap for collecting newly emerged stoneflies (Plecoptera). Oikos 31:323-325. Merritt, Richard W.; Cummins, Kenneth W. 1984. An introduction to the aquatic insects of North America (2nd ed.). Kendall/Hunt Publ. Co., Dubuque, Iowa. 722 p. Minshall, G. Wayne.; Petersen, Robert C.; Cummins, Kenneth W.; Bott, Thomas L.; Sedell, James R.; Cushing, Colbert E.; Vannote, Robin L. 1983. Interbiome comparison of stream ecosystem dynamics. Ecological Monographs 53:1-25. Minshall, G. Wayne; Cummins, Kenneth W.; Petersen, Robert C.; Cushing, Colbert E.; Bruns, Dale A.; Sedell,

James R.; Vannote, Robin L. 1985. Developments in stream ecosystem theory. Canadian Journal of Fisheries and Aquatic Sciences 42:1045-1055. Molles, Manuel C., Jr. 1982. Trichopteran communities of streams associated with aspen and conifer forests: longterm structural change. Ecology 63:1-6. Ross, Herbert H. 1963. Stream communities and terrestrial biomes. Arch. Hydrobiol. 59:235-242. Stark, Bill P.; Szczytko, Stanley W.; Baumann, Richard W. 1986. North American stoneflies (Plecoptera): systematics, distribution, and taxonomic references. Great Basin Naturalist 46:383-397. Statzner, Bernhard; Higler, Bert. 1985. Questions and comments on the River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences 42:1038-1044. Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology. Transactions of the American Geophysical Union 38:913-920. Vannote, Robin L.; Minshall, G. Wayne; Cummins Kenneth W.; Sedell, James R.; Cushing, Colbert E. 1980. The River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences 37:130-137. Wiggins, Glenn B.; Mackay, Rosemary J. 1978. Some relationships between systematics and trophic ecology in Nearctic aquatic insects, with special reference to Trichoptera. Ecology 59:1211-1220. Winterbourn, M.J.; Rounick, J.S.; Cowie, B. 1981. Are New Zealand stream ecosystems really different? New Zealand Journal of Marine and Freshwater Research 15:321-328. Zwick, Peter. 1973. Insecta: Plecoptera. Phylogenetisches System and Katalog. Das Tierreich 94:1-465.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

241

HABITAT AND POPULATIONS OF THE VALLEY ELDERBERRY LONGHORN BEETLE ALONG THE SACRAMENTO RIVER 1
F. Jordan Lang, James D. Jokerst and Gregory E. Sutter2 Abstract: Prior to 1985, the valley elderberry longhorn beetle, a threatened species protected under the federal Endangered Species Act, was known only from northern California riparian areas along the American River and Putah Creek in the Sacramento Valley, and along several rivers in the northern San Joaquin Valley. During 19851987, our study extended the known range of the beetle northward along the Sacramento River to Red Bluff by searching riparian areas for elderberry shrubs, the required habitat for beetle larvae; by searching shrubs for beetle emergence holes; and by searching for adult beetles during their brief spring emergence period. The survey results showed that the beetle is distributed widely along the river but is rare. The valley elderberry longhorn beetle (Democerus californicus dimorphus Fisher [Coleoptera: Cerambycidae]), herein abbreviated VELB, is federally designated as a threatened species under the Endangered Species Act. VELB is endemic to riparian areas in the Sacramento and San Joaquin Valleys of California, where its larvae inhabit elderberry (Sambucus sp.) shrubs. Because of the severe reduction in the Central Valley's natural riparian vegetation over the last 150 years, the beetle's habitat has become scarce and discontinuous. This paper describes a study to determine (a) locations of active colonies of the VELB along the Sacramento River from the City of Sacramento upstream to Red Bluff, and (b) the distribution of potential and actual VELB habitat within the same reach. Potential VELB habitat was defined by the presence of elderberry shrubs within areas of riparian vegetation. Actual VELB habitat was defined by the presence of VELB emergence holes in elderberry shrubs. Active colonies were defined by detection of adult VELB during their spring emergence period in April and May. nia , from Los Angeles northward to Mendocino County. The two subspecies occupy distinct geographic ranges (Linsley and Chemsak 1972). VELB (i.e., D. c. dimorphus) are moderate-sized, brightly colored, sexually dimorphic beetles with female body lengths of 18-25 mm and male body lengths of 1321 mm (Linsley and Chemsak 1972). Their antennae are long, extending more than half their body lengths. Males exhibit one of two different color patterns. In some, the elytra (i.e., the first pair of hardened wings that protect the flight wings) resemble those of the coastal subspecies (D. c. californicus), whose elytra are dark metallic green with bright reddish orange borders (U.S. Fish and Wildlife Service 1984). In other VELB males, the solid green pattern is reduced to four oblong spots on the elytra. Intergrades between the two divergent color patterns also exist. In general, both male and female VELB are smaller than D. c. californicus. Little is known about the life history of VELB; it is assumed, however, to follow a sequence of events similar to that of several related species of Desmocerus beetles (U.S. Fish and Wildlife Service 1984). Females may lay as many as 200 eggs during their short life span. Eggs are laid on foliage or leaf petioles at the outer tips of branches of living elderberry plants and in bark crevices of larger stems and trunks. Presumably, the eggs hatch 2-3 days after they are laid. Larvae bore through the pith of stems and possibly through the roots of elderberry shrubs. When larvae are ready to pupate, they work their way through the pith, open an emergence hole through the bark, and then return to the pith for pupation. Adults exit through the emergence holes, approximately 0.5 cm in diameter, and then can be found on elderberry foliage, flowers, or stems, or on associated plants (U.S. Fish and Wildlife Service 1984). VELB was first described by Fisher in 1921 from specimens collected in Sacramento (U.S. Fish and Wildlife Service 1984). By 1984, VELB was known from only three Central Valley drainages: Merced River, Putah Creek, and the American River. Since designation of the VELB as threatened in 1980, a number of field surveys (in addition to those reported here) have produced new records of VELB through observations of adult beetles or their emergence holes. These new locations include many sites on the American River floodplain; the

The Valley Elderberry Longhorn Beetle


Two subspecies of elderberry longhorn beetles, (D. c. dimorphus and D. c. californicus), occur in California. The California elderberry longhorn beetle (D. c. californicus) occupies valleys of the coastal ranges of Califor1

Presented at the California Riparian Systems Conference; September 22-24, 1988; Davis, California. Environmental Scientist, Environmental Specialist, and Environmental Specialist, respectively, Jones & Stokes Associates, Inc., Sacramento, California. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

242

Cosumnes River floodplain; sites along several creeks near Roseville, California, west of Folsom Lake; and sites along the Mokelumne, Calaveras, San Joaquin, Middle, Tuolumne, and Merced Rivers in the San Joaquin Valley (Jones & Stokes Associates, Inc. 1987).

Study Area
The study area encompassed the floodplain of the Sacramento River, from its confluence with the American River at River Mile (RM) 60 in the City of Sacramento to Red Bluff Diversion Dam at RM 243. Within this 183-mile-long reach, the study focused on areas occupied by native riparian vegetation, which comprises a mosaic of riparian scrub on levees and other regularly disturbed areas, young-growth cotton wood-willow forest on lower terraces, and mature mixed-riparian forest on higher terraces. Many intervening sites have been cleared for agriculture or installation of levee protection. Other papers presented in these proceedings, such as by McCarten (1988), describe riparian vegetation along the Sacramento River in greater detail.

Spring surveys for adult VELB were conducted in stands of elderberries with emergence holes previously detected during the winter surveys. A three- or fourperson crew walked slowly through the elderberry stands, searching the leaves, flowers, and stems of each shrub, sometimes using binoculars. Spring surveys were conducted during the late April to late May period when adult beetles emerge from their larval galleries. Where adult beetles were collected, data were taken on characteristics of the VELB specimens, the substrates on which they were found, weather conditions, elderberry phenology, and vegetation structure. All collected VELB specimens were deposited in the University of California, Berkeley, Essig Museum.

Results

Elderberries as Potential Habitat The winter surveys over 3 study years (i.e., 1985, 1986, and 1987) covered a total of 1,943 hectares in 66 polygons of riparian vegetation (Table 1). Average sizes of surveyed polygons varied among three sections of the study area. In the Sacramento-Colusa reach, 25 polygons averaged about 15 hectares per polygon; in the Colusa-Chico Landing reach, 23 polygons averaged 30 hectares per polygon; and in the Chico LandingRed Bluff reach, 18 polygons averaged 50 hectares per polygon. This pattern reflects the fact that riparian vegetation below Colusa is confined to narrow strips within tightly constraining levees. From Colusa to Chico Landing, setback levees encompass large areas of riparian vegetation, and above Chico Landing, the absence of levees is associated with the largest stands of riparian vegetation. Sixty-three of the 66 polygons surveyed throughout the study contained stands of elderberry shrubs (Table 1). Two of the three polygons without elderberries were located in the Sacramento-Colusa reach, possibly reflecting higher intensities of maintenance on narrow berms within the levees below Colusa.

Methods
Riparian vegetation within the study area was delineated on aerial photographic map plates in the Sacramento River Aerial Atlas of the U.S. Army Corps of Engineers (U.S. Army Corps of Engineers 1984). At least 30 percent of the riparian vegetation area was selected randomly and visited during the winter surveys. Individual survey polygons were selected so that the probability of selection was proportional to the size of the polygon of delineated riparian vegetation (Husch and others 1972). Polygons of riparian vegetation varied widely in size from less than 1 hectare to more than 50 hectares. Use of the proportional probability selection procedure ensured that large-area polygons would not be underrepresented in the survey. Each selected riparian polygon was searched thoroughly on the ground to determine the presence or absence of elderberry shrubs. If present, elderberry shrubs were in turn thoroughly searched for VELB emergence holes by inspecting the shrub stems carefully. Holes were presumed to have been produced by VELB emergence if they were of the appropriate shape and size (i.e., about 0.6 cm diameter and uniformly cylindrical). No other Cerambycid borers are known to occupy elderberry shrubs in Central Valley riparian areas. Stem searches were often quite difficult because of dense understories of poison-oak and blackberries. Any elderberries found to have VELB emergence holes were characterized on field data sheets.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

VELB Emergence Holes VELB emergence holes were found in 42 (64 percent) of the 66 polygons covered by the winter surveys (Table 1). Percentages of polygons with detected emergence holes increased from 28 percent below Colusa to 94 percent above Chico Landing. Evidently, habitat use by VELB has been wide-spread throughout the upstream portions of the study area. 243

Table 1

- Aggregate survey results for 3 study years.


Specimen Characteristics Height Activity During During Observation Sex Observation (c) Male Male Female Male Female Male Male Female Male Female Walking, flying Walking Resting Flying, resting Resting Resting Resting Resting Walking Resting 2.1 1.5 0.9 2.1 2.1 1.8 3.0 2.4 1.2 0.08 Substrate Characteristics Stem Total Exposure Diameter Height of (m) (m) Shrub Overstory 7.6 5.1 2.5 5.1 10.2 1.9 14.0 10.2 10.2 11.4 6.0 2.0 1.2 4.6 3.0 3.0 6.0 6.0 4.6 4.6 Sun Sun Part shade Sun Shade Shade Sun Sun Sun Sun Open Open Open Open Mixed riparian Mixed riparian Open Open Open Open

River Mile 84.3 West 84.3 West 126.5 West 235 West 138.7 West 138.7 West 169.5 West 177.7 East 179.4 East 179.4 East

Collection Date 5/3/85 5/3/85 5/3/85 5/5/8S 4/29/87 4/29/87 4/29/87 5/7/87 5/8/87 5/8/87

Time of Day 1200 1226 1558 1105 1100 1130 1540 1045 1022 1350

Weather Clear, gentle breeze Clear, gentle breeze Clear, no wind Clear, no wind 20% clouds, no wind 20% clouds, no wind 50% clouds, slight breeze 20% clouds, no wind 20% clouds, no wind 30% clouds, no wind

Temp. 77 F 77 F 75 F 78 F 70 F 75 F 80 F 85 F 80 F 90 F

Stem Diameter at VELB Emergence Holes Stem diameters at 49 current-year emergence holes detected in May 1987 are shown by size classes in Figure 1. Current-year holes were positively identified by the presence of fresh wood and frass (i.e., wood shavings usually mixed with excrement produced by the boring insect). No holes were detected on stems less than 2.5 cm in diameter, whereas the largest stem with a hole was 20 cm in diameter. Sixty-six percent of the holes were observed in stems less than 7.0 cm in diameter, and 34 percent were observed in stems greater than this size.

Stem Height at VELB Emergence Holes Stem heights were estimated at 627 VELB emergence holes observed during 1986 and 1987 surveys. Nearly 70 percent of the observed emergence holes were at heights at or below 1.2 m on elderberry stems (Figure 2). Approximately 10 percent were detected at heights above 1.8 m.

Figure 1- Frequency distribution of stem diameters at current-year holes detected in May 1987.

Figure 2- Frequency distribution of stem heights at VELB exit holes detected in 1986 and 1987.

244

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Table 2

Descriptions of collections of adult VELB specimens.


Winter Surveys Spring Surveys Area Surveyed (ha) 40 Polygons Surveyed 4 Polygons with Adult VELB No. Percent 3 75

Reach

Area Surveyed (ha) 383

Polygons Surveyed 25

Polygons with: Elderberries VELB Exit Holes No. Percent No. Percent 23 92 7 28

SacramentoColusa (RM 60-145) Colusa Chico Landing (RM 145-195) Chico LandingRed Bluff (RM 195-243) TOTAL

688

23

22

96

18

78

523

16

19

875

18

18

100

17

94

274

13

1,946

66

63

95

42

64

837

33

21

Characteristics of Elderberry Clumps with VELB Exit Holes

A total of 367 elderberry clumps (i.e., groups of stems apparently arising from the same root stock) with VELB emergence holes were observed during the three winter surveys. Elderberry clumps with holes were found in young-growth riparian stands of young cottonwoods and willows on low-terrace areas of the Sacramento River floodplain (15 percent of 360 clumps); riparian stands of mature and senescent cottonwoods, generally on lowterrace areas, often with woody vines of wild grape (35 percent); mature riparian stands of mixed tree species including cottonwood, boxelder, walnut, or valley oak, often with woody vines of wild grape, on high-terrace areas (35 percent); and both high- and low-terrace areas with no overstory (14 percent of clumps). Adult VELB Collections

spring 1986, when no VELB were found anywhere in the study area, possibly due to effects of the February 1986 flooding along the Sacramento River, unusual weather conditions, and patterns of elderberry phenology during spring 1986. Above Chico Landing, adult VELB were collected in only 1 of the 13 polygons (Table 1). The low collection rate here also reflected the regionwide lack of collections in 1986. More details on study results can be found in the report by Jones & Stokes Associates (1987).

Discussion

VELB Rarity Ten VELB adults were collected at seven different locations on five different dates in spring 1985 and spring 1987 (Table 2). The adult beetles were collected in seven of the 33 polygons surveyed during spring (Table 1). Three of the four polygons below Colusa provided VELB collections, indicating the importance of narrow riparian strips with elderberries as active VELB habitat. Adult VELB were collected in 3 of the 16 polygons surveyed during spring between Colusa and Chico Landing. Nine of these 16 polygons were surveyed during late
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

During this study, approximately 70 person-days of field time were spent searching habitat containing VELB emergence holes. This effort resulted in collections of only 10 adult VELB, showing that VELB is a rare insect. Our low rate of collection may have been influenced by two other factors in addition to the species' rarity. Visual searches for VELB adults are difficult within suitable habitat because elderberry shrubs are tall and because VELB adults are generally more active in higher portions of shrubs (e.g., 5 of 10 adults were 245

collected at heights of 2 m or more). Many shrubs are inaccessible because of dense understories or overhanging vines of wild grape, blackberry, and poison-oak. In addition, VELB adults are not greatly mobile, they do not concentrate on elderberry flowers, and they do not often fly about, even in hot weather.

emergence holes. Thus, the VELB population along the Sacramento River may be larger than is currently indicated by frequencies of emergence holes. Sizes and Ages of Elderberries Used by VELB Nearly 70 percent of current-year exit holes found in May 1987 were in stems 2.5-6.5 cm in diameter at the emergence hole. None were in stems less than 2.5 cm in diameter. Because of the rapid growth rate of young elderberries, stems less than about 6 cm in diameter may have grown from new shoots within the 2-year period of VELB larval occupation. The best strategy for VELB habitat management is to maintain mixed stands containing all ages and sizes of elderberries. Our data suggest that VELB larvae are able to inhabit a broad range of ages and sizes of elderberries over a minimum size threshold of about 2.5 cm basal stem diameter. Where VELB habitat restoration or enhancement is being considered, a mixture of sizes of elderberries should be established to increase habitat area and quality.

VELB Distribution on Sacramento River Floodplain VELB emergence holes were detected in elderberries in most riparian stands surveyed, indicating that VELB are widespread, although rare, in suitable riparian habitat along the Sacramento River. The relatively small proportion of current-year emergence holes suggests that the population is limited at any one site by factors other than habitat availability. Within the Sacramento River floodplain, VELB does not appear to be restricted to particular kinds of riparian vegetation or floodplain topography. VELB emergence holes were found in nearly all situations, ranging from isolated elderberry clumps in savannalike areas to continuous stands beneath tall overstories, areas with or without extensive woody understory vegetation, and on both low- and high-terrace floodplains. VELB requires established elderberry plants of mature size and age. No emergence holes were found on stems less than 2.5 cm in diameter, consistent with the presumed 2-year period that VELB larvae are thought to inhabit elderberry shrubs. Stems of less than 2.5 cm in diameter may be inhabited by VELB larvae but grow greatly in diameter within the 2-year period. Annual cutting, burning, or herbicide treatment of elderberry shrubs would preclude VELB occupancy.

Conclusions
The currently known range of VELB extends through the San Joaquin- Sacramento Valley from the Merced River in Merced County northward to near Red Bluff on the Sacramento River in Tehama County. Our study extended the known range northward by about 120 air miles and 183 river miles, from the City of Sacramento to Red Bluff. Other recent studies have extended the known distribution to at least 6 rivers in the San Joaquin Valley and along the American River through the City of Sacramento. The currently known range is based on observations of either adult VELB or VELB emergence holes in elderberry shrubs. In our winter surveys along the Sacramento River, 95 percent of all surveyed stands of riparian vegetation contained elderberry shrubs, and VELB emergence holes were detected in 64 percent of the surveyed stands. During our spring surveys in stands with emergence holes, we collected adult VELB in 7 of the 34 stands surveyed.

Role of Birds If birds forage after VELB larvae, they may be an important factor regulating VELB populations. Bird holes were found at nearly every site where VELB emergence holes were found. A key need for monitoring VELB populations is to determine the role of birds in creating holes in elderberry stems, in enlarging or reworking VELB emergence holes, and in preying on VELB larvae. Easily distinguished bird holes were at least as common as VELB holes at many sites and were far more abundant than VELB holes at some sites. We dissected several elderberry stems with bird holes and found evidence of tunneling by insect larvae, presumably VELB, in the pith. In searching for evidence of VELB, we recorded only clearly identifiable Cerambycid emergence holes and rejected holes that appeared to be created or enlarged by birds. If birds commonly enlarge holes originally bored by emerging VELB, they would obscure them as VELB 246

Acknowledgments
We thank Jack E. Williams, U.S. Fish and Wildlife Service, Sacramento Endangered Species Office; William Shepard, California State University, Sacramento; Paul A. Rude; A. Miriam Green; and Clinton Kellner for contributing substantially to the success of this study.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

The study was supported by U.S. Fish and Wildlife Service Contract No. 14-16-0001-84279(NR).

Linsley, E. G.; Chemsak, J. A. 1972. Cerambycidae of North America, Part VI, No. 1. Taxonomy and classification of the subfamily Lepturinae. University of California Publications in Entomology 69:1-13. McCarten, N. 1989. Plant community development, site quality analysis and river dynamics in the design of riparian preserves on the middle Sacramento River, California, [these proceedings] U.S. Army Corps of Engineers, Sacramento District. 1984. Sacramento River, California, aerial atlas. Investigations Section A, Water Resources Planning Branch. Sacramento, CA. 41 plates. U.S. Fish and Wildlife Service. 1984. Valley elderberry longhorn beetle recovery plan. Portland, OR. 61 p.

References
Husch, B.; Miller, C. I.; and Beers, T. W. 1972. Forest mensuration. 2nd edition. New York, NY: The Ronald Press Company; 410 p. Jones & Stokes Associates., Inc. 1987. Final report-survey of habitat and populations of the valley elderberry longhorn beetle along the Sacramento River. Prepared for: U.S. Department of the Interior, Fish and Wildlife Service, Region 1, Endangered Species Office, Sacramento, CA. 48 p.

USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

247

PRACTICAL TECHNIQUES FOR VALLEY ELDERBERRY LONGHORN BEETLE MITIGATION 1


Greg Sutter, Jeurel Singleton, Jim King, and Ann Fisher The valley elderberry longhorn beetle (Desmocerus californicus dimorphus) is a live-wood pith-boring beetle of the family Cerambycidae. Elderberry (Sambucus spp.) shrubs are the beetle's exclusive host plant. The valley elderberry longhorn beetle (VELB) is classified as a threatened species by the U.S. Fish and Wildlife Service. The VELB is sporadically distributed in the Sacramento and San Joaquin Valleys from Red Bluff to Fresno, California. This poster presentation describes mitigation measures for VELB habitat loss that has resulted from public and private development at sites along the Sacramento and American Rivers.
2

Foliage and stems should be pruned to reduce the shoot:root ratio and thus reduce transpirational load until the root system reestablishes. In addition, transplanted shrubs should be irrigated on a schedule that will gradually wean them from artificial moisture supplies. The large-diameter plant material pruned from mature elderberries also should be moved to the transplant site because larvae inhabiting the prunings may emerge as VELB adults during the spring emergence period and lay eggs on the transplanted elderberries. As mitigation for a large land development project, mature elderberry plants were moved by Jones & Stokes Associates to a parcel deeded to the county as an addition to the American River Parkway. Approximately 80% of the transplanted plants were surviving after the first growing season.

Mitigation Techniques and Project Examples


When designing a mitigation program for VELB, it is essential to consider the ecological characteristics of both the VELB and the elderberry host plants, site conditions at both the impact site and the mitigation site if offsite mitigation is used, and the importance of proper timing when dealing with natural systems. The best way to mitigate project impacts is to build protection of existing elderberry habitat into the project. This is difficult in urban settings where public use and activities reduce the potential for long-term survival and regeneration of elderberry shrubs.

Revegetating with Seedlings or Rooted Cuttings Elderberry seedlings or rooted cuttings also can be used for offsite habitat enhancement or compensation. An ideal offsite mitigation plan would use young plants in combination with transplanted mature plants. The easiest way to grow young plants for revegetation is from seed. If seed is not available, cuttings can be taken either in the dormant season as hardwood cuttings or in the spring as softwood cuttings. Rooted cuttings planted by the California Department of Water Resources in the American River Parkway as mitigation for a floodway vegetation removal program are in their second growing season after planting. The plants were irrigated regularly through the first growing season (every 10 days to 3 weeks). The watering was reduced in the second season, and no watering will occur in following years. The plants had a high survival rate of about 90% after the first growing season, and at the end of the second year well over 50% are still thriving.

Transplanting Mature Shrubs Where preservation of habitat in place is not possible, mitigation programs can include habitat enhancement offsite. Transplanting mature elderberries occupied by VELB to a new site is one way to mitigate by offsite habitat enhancement. Unlike offsite mitigation using elderberry seedlings or rooted cuttings, the transplanting of mature plants, if successful, can also translocate VELB larvae within stems to the new site. Successful transplanting of mature elderberry plants when the plants are dormant (approximately November through the first two weeks of February)
1Presented 2

at the California Riparian Systems Conference, September 22-24, 1988, Davis, California

Resource Ecologist, Jones & Stokes Associates Inc., Sacramento, Calif.; Entomologist, U.S. Fish and Wildlife Service, Sacramento Endangered Species Office, Sacramento Calif.; Environmental Specialist, The California State Resources Agency, Department of Water Resources, Division of Flood Management, Sacramento Calif.; Owner and Manager, Cornflower Farms, Elk Grove, Calif. USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

248

Figure 1- Adult VELB emergence hole in an elderberry stem.

Figure 2- A mature elderberry plant heavily pruned for transplanting with many new sprouts after planting.

Figure 3- A transplanted elderberry shrub with vigorous new growth. Notice the large watering basin.
USDA Forest Service Gen. Tech. Rep. PSW-110. 1989.

Figure 4- Rooted elderberry cuttings ready to be planted. 249

HOW TIGHT IS THE LINKAGE BETWEEN TREES AND TROUT? 1


Margaret A. Wilzbach 2 Abstract: This paper explores the tightness of the linkage between stream-dwelling salmonids and riparian vegetation. Comparison of original distributions of salmonid species with that of vegetation types shows that distribution within a given salmonid species is not limited to a specific vegetation type, and that different salmonid species co-occur within a given vegetation type. Examination of reported differences in trout production among streams appear to be related to differences in riparian setting only indirectly and insofar as these reflect differences in prey availability and, to a lesser extent, differences in habitat features. Variability in trout production estimates are minimized when comparisons are species-specific and normalized for temperature differences among streams. Within a riparian vegetation type, the relationship between trout production and successional age of the streamside vegetation is often inverse.

Native Distributions of Salmonids in Relation to Vegetation


The native ran